Molecular biology of doxorubicin-induced cardiomyopathy

Abstract

The anthracycline doxorubicin is an antineoplastic agent, eliciting chronic cardiac toxicity. It occurs in patients after prolonged administration of doxorubicin, leading to congestive heart failure. The pathogenesis of the doxorubicin-induced car-diomyopathy is not well understood. The present article summarizes the unique effect of doxorubicin on cardiac-specific gene expression. In addition to binding to DNA, doxorubicin directly affects the function of a variety of proteins. Free radical generation, damage to mitochondria and active cell death are also critical in the development of doxorubicin-induced cardiac toxicity. Agents providing effective cardioprotection are also reviewed.

The anthracycline doxorubicin is a widely used antineoplastic agent, active against a variety of neoplasms. A severe, cumulative, dose-dependent chronic cardiac toxicity is the major limitation of anthracycline therapy. Chronic cardiotoxicity occurs in patients after prolonged administration of doxorubicin; a similar cardiotoxicity can be elicited in many animal species, including mouse, rat, rabbit, dog and monkey, after treatment with doxorubicin. The cardiotoxicity consists of a chronic, progressive cardiomyopathy with myocyte vacuolation and degeneration, interstitial edema and fibroplasia leading to congestive heart failure. Despite considerable work on the subject, the pathogenesis of doxorubicin-induced cardiomyopathy is not well understood. However, doxorubicin has been shown to exert a multiplicity of complex biochemical effects on the myocardium, including the following: binding to DNA and alteration of nucleic acids and protein synthesis, lipid peroxidation after free radical generation, release of histamine and catecholamines, damage to mitochondria, an effect on various cellular membranes, excess calcium influx and an effect on collagen matrix. A combination of these effects probably triggers the myocardial lesion.

DOXORUBICIN DISRUPTS THE CARDIAC-SPECIFIC PROGRAM OF GENE EXPRESSION DIRECTLY OR THROUGH FREE RADICALS

Recent experiments have shown that doxorubicin has a unique effect on cardiac-specific gene expression. A constitutively expressed, cardiac-restricted nuclear protein whose mRNA level is exquisitely sensitive to doxorubicin was identified in cloned neonatal cardiac myocytes by the differential display strategy. Cardiac adriamycin-responsive protein (CARP) mRNA was present at the earliest stages of cardiac morphogenesis. CARP appeared to function as a negative regulator of cardiac-specific gene expression. Overexpression of CARP in cardiomyocytes suppressed cardiac troponin C and atrial natriuretic factor transcription. Cotransfection experiments in HeLa cells indicated that CARP inhibited Nkx2.5 transactivation of atrial natriuretic factor promoter. When fused to a GAL4 DNA-binding domain, CARP had transcriptional inhibitory properties in noncardiac cells. CARP is the first example of a cardiac-restricted transcriptional regulatory protein that is sensitive to doxorubicin (1).

Doxorubicin also directly affects the cardiac-specific gene expression of protein-regulating energy and calcium metabolism (2). Transcripts of important nuclear genes encoding enzymes critical in production of energy in cardiomyocytes – ADP/ATP translocase, a heart- and muscle-specific isoform; Reiske iron-sulphur protein, a ubiquitously expressed electron transport chain component; and a muscle isozyme of phosphofructokinase – are rapidly degraded. Loss of these mRNAs is evident as early as 2 h after doxorubicin administration and precedes significant reduction of intracellular ATP. Loss of mRNAs generating ATP was highly selective because mRNAs for other energy production enzymes (cytochrome c, cytochrome b and malate dehydrogenase) and genes important in glycolysis (pyruvate kinase and glyceraldehyde-3-phosphate dehydrogenase) were unaffected even at 24 and 48 h. These findings may link doxorubicin-induced damage to membranes and signalling pathways with, first, suppression of transcripts encoding myofibrillar proteins and proteins of energy production pathways and, second, depletion of intracellular ATP stores, myofibrillar degeneration and related cardiotoxic effects (3).

Expression of genes encoding proteins that affect calcium ion homeostasis was altered in the hearts of rabbits chronically treated with doxorubicin (2.5 mg/kg intravenously, eight weeks after the final injection). Cardiac output was significantly decreased in the doxorubicin-treated rabbits. Concentrations of mRNA for the sarcoplasmic reticulum proteins were significantly diminished in the doxorubicin-treated hearts: ryanodine receptor-2, sarcoplasmic reticulum Ca²⁺-ATPase and phospholamban cardiac calsequestrin. The amount of sarcoplasmic reticulum Ca²⁺-ATPase protein and the calcium uptake capacity were concomitantly decreased with the level of expression of its mRNA. These findings suggest that a selective decrease in mRNA expression of sarcoplasmic reticulum Ca²⁺ transport proteins is responsible for the impaired calcium handling and, thus, for the reduced cardiac function seen in the cardiomyopathy induced in rabbits by long term treatment with doxorubicin (4).

The role of free radicals in the genesis of doxorubicin-induced chronic cardiotoxicity and the cardioprotective effects of the spin trap N-tert-butyl-alpha-phenylnitrone (PBN) were investigated in an in vivo rat model. Because doxorubicin and free radicals are no longer present in the myocardium by the time the delayed effects of the drug become apparent, doxorubicin has been proposed to act by causing early radical-dependent DNA lesions, resulting in impaired synthesis of critical target proteins. DNA lesions were detected 10 days after doxorubicin treatment (3×3 mg/kg intravenously) and were still present at the onset of chronic dilated cardiomyopathy. PBN, administered throughout the time of persistence of doxorubicin in the myocardium (approximately two weeks), prevented the development of DNA lesions, as well as the late contractile and electrical impairment induced by anthracycline, thus supporting the hypothesis that free radicals play a causal part in both phenomena. Moreover, a deletion of about 4 kb has been found in the mutated mitochondrial DNA in cardiomyocytes with chronic doxorubicin-induced cardiotoxicity in mouse. The incidence of the mutated DNA deletion increased with the dosage and with the duration of doxorubicin administration (5,6).

ACUTE AND CHRONIC EFFECTS OF DOXORUBICIN ON CARDIAC METABOLISM AND THE CYTOSKELETON

Besides the critical role of doxorubicin in the generation of mutation and the alteration of cardiac-specific gene expression, it also directly affects the function of a variety of proteins. There are important clues showing that doxorubicin interferes with the oxidative metabolism of myocardium. First, doxorubicin induced significant changes in the activity of the oxidation-sensitive enzyme creatine kinase in the heart in vivo and in a cardiomyocyte culture model (7). Second, acute exposure to adriamycin caused a concentration- and time-dependent inhibition of carnitine palmitoyl transferase-1 dependent long chain fatty acid (palmitate) oxidation. After in vitro or in vivo administration, adriamycin inhibited fatty acid oxidation in part secondary to inhibition of carnitine palmitoyl transferase-1 and to depletion of its substrate, l-carnitine, in cardiac tissue (8).

The integrity of contractile proteins has also been jeopardized by doxorubicin administration. In vivo atrial natriuretic factor and beta-myosin heavy chain expression increased with dose and time after single and multiple doxorubicin injections. Moreover, the release of cardiac troponin T (cTnT) as a biomarker of doxorubicin-induced chronic cardiac injury was evaluated in the spontaneously hypertensive rat (SHR) model. Increased serum concentrations of cTnT and decreased immunohistochemical staining of heart sections for this protein were noted in SHRs treated with cumulative doses of doxorubicin (7 mg/kg) that induced only minimal histological alterations in myocytes. Concentrations of cTnT were further increased, coincident with reduced immunohistochemical staining, in SHRs given 10 to 12 mg/kg doxorubicin. Thus, monitoring serum concentrations of cTnT can detect doxorubicin-induced myocyte damage in SHRs and may prove useful for the noninvasive evaluation of this toxicity in humans (9).

Troponins proved to be highly cardiospecific markers with remarkable diagnostic windows (increased serum concentrations are detectable for a longer period) and their detection is clinically convenient. They have been incorporated into diagnostic procedures in cardiology for prevention as well as prognosis. It is important to note that the evaluation of TnT directly influences the therapeutic intervention. Animal models of cardiotoxicity of chemotherapeutics have significantly influenced the development of clinical oncology. Analysis of troponin concentrations in animal models of chemotherapeutically induced cardiotoxicity was followed by a variety of clinical trials and studies. The outcomes of clinical studies were not identical: the diversity of results was caused by differences in methods (diagnostic limits, cut-off), different cumulative doses of anthracyclines, small numbers of patients enrolled, a variety of chemoptherapeutic regimens and different techniques for evaluating left ventricular function. In summary, troponins may serve as an important predictor of the subclinical phase of the cardiotoxic effect of anthracyclines and other chemotherapeutics.

CRITICAL ROLE OF MITOCHONDRIA IN TRIGGERING DOXORUBICIN CARDIOTOXICITY

In a complete understanding of the organ-specific mechanism of doxorubicin-induced cardiotoxicity, mitochondria are unequivocally accepted as the locus where the molecular disorder is rapidly triggered. Isolated cells from adult rat hearts preloaded with 2′,7′-dichlorofluorescein (an oxidant-sensitive fluorescent probe) and exposed to doxorubicin (40 to 160 μM) showed intracellular oxidation close to the mitochondria after only 20 min (10). A growing number of reports intimate that unbalanced oxygen activation is established through heart mitochondria in the presence of anthraquinones. In contrast to liver mitochondria, isolated heart mitochondria have been unequivocally shown to shuttle single electrons to anthraquinones, giving rise to superoxide anion radical formation by auto-oxidizing anthraquinone semiquinones. The involvement of exogenous NADH dehydrogenase in this deleterious electron deviation from the respiratory chain was also shown. This enzyme that is associated with complex I of the respiratory chain catalyzes the oxidation of cytosolic NADH. Anthraquinone activation through isolated heart mitochondria was reported to require the external addition of NADH, suggesting a flux of reducing equivalents from NADH to anthraquinones in the cytosol. Unlike heart mitochondria, intact liver mitochondria, which are lacking this NADH-related pathway of reducing equivalents from the cytosol to the respiratory chain, cannot be made to activate anthraquinones to semiquinones by NADH or any other substrate of respiration. It appears, therefore, that the exogenous NADH dehydrogenase of heart mitochondria exerts a key function in the myocardial toxicogenesis of anthraquinones through oxygen activation through semi-reduced anthraquinones. Assessing the toxicological significance of the exogenous NADH dehydrogenase in anthraquinone-related heart injury requires analysis of reaction products and their impact on vital bioenergetic functions, such as energy gain from the oxidation of respiratory substrates. The identity and possible interactions of radical species emerging from NADH-respiring heart mitochondria in the presence of anthraquinones were analyzed by electron spin resonance technique. The following metabolic steps occurred, causing depression of energy metabolism in the cardiac tissue. After one-electron transfer to the parent hydrophilic anthraquinone molecule, destabilization of the radical caused cleavage of the sugar residue. Accumulation of the lipophilic aglycone metabolite in the inner mitochondrial membrane diverted electrons from the regular pathway to electron acceptors out of sequence such as hydrogen peroxide. Hydroxy radicals are formed and affect the functional integrity of energy-linked respiration. The key and possibly initiating role of exogenous NADH dehydrogenase of cardiac mitochondria in this reaction pathway provides a rationale to explain the selective cardiotoxic potency of the cytostatic anthraquinone glycosides (11).

Additionally, mitochondrial calcium plays an important part in adriamycin-induced cardiotoxicity. Drugs such as cyclosporine A or tacrolimus (FK506) alter the permeability of heart inner mitochondrial membrane and oppose the cardiotoxicity. The ability of heart mitochondria isolated from adriamycin-treated rats to retain accumulated calcium was sharply reduced. The increase of diagnostic serum enzymes and isoenzymes and the reduced ability of heart mitochondria to retain calcium was restored to almost the normal levels when (500 μg/kg body weight) of cyclosporine A or FK506 was injected with adriamycin. The data suggested that adriamycin cardiotoxicity may be caused by the increased inner membrane permeability in heart mitochondria as a result of increasing the sensitivity of a calcium-dependent pore of the inner mitochondrial membrane to calcium, leading to dissipation of membrane potential and release of preaccumulated calcium. Similarly, impaired sequestration of intracellular free calcium ions in individual myocytes may be one factor leading to diastolic dysfunction. Monitoring diastolic function is important to detect early cardiotoxicity caused by doxorubicin (2). Suitable antagonists of calcium-dependent pore formation such as cyclosporine A or FK506 may improve heart tolerance to adriamycin (12).

FREE RADICALS TRIGGER DOXORUBICIN CARDIOMYOPATHY

Doxorubicin induced a cascade of early biochemical changes characterized by the presence of aldehydic lipid peroxidation products. The results showed that aldehydes in rat plasma and heart tissues increased significantly following doxorubicin treatment (13). The changes occurred early and peaked about 2 h after doxorubicin administration. Toxic aldehyde concentrations including malondialdehyde, hexanal and 4-hydroxy-non-2-enal glutathione concentrations decreased early. Thus, the experimental data confirmed the involvement of free radicals and suggested that the cytotoxic aldehydes play a central part in initiating the steps that lead to functional impairment of the myocardium following doxorubicin administration. Scavengers and metabolic removal of some of the aldehydes also play a part in protecting the myocardium against injury. Cardiomyocytes with reduced superoxide dismutase activity, but normal ATP content and viability, were obtained by treating isolated cells with diethyldithiocarbamate (DDC). DDC-treated myocytes were significantly less resistant to doxorubicin than controls. Doxorubicin-stimulated super-oxide anion formation, measured by the rate of superoxide dismutase-inhibitable acetylated cytochrome c reduction, was significantly higher in the cytosolic fraction of DDC-treated cells than in controls. These results indicate that for isolated cardiac myocytes an essential part of cytotoxicity of doxorubicin can be explained by the formation of superoxide anion and that the level of intracellular superoxide dismutase activity should be considered to be a significant factor in cell protection (14,15).

DOXORUBICIN-INDUCED CELL DEATH OF CARDIAC MYOCYTES

In vitro experiments, as well as data from animal models, refer to a potential role of active cell death in doxorubicin-induced cardiotoxicity. Cardiomyocytes isolated from rat hearts were exposed to 20 μM adriamycin for 1 h and examined at a range of post-treatment durations (0 to 23 h). Adriamycin caused a significant decrease in rod-shaped cells and an increase in round cells. Both Hoechst 33258 staining and terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling assay showed a significantly increased number of apoptotic myocytes and nucleosomal fragmentation on exposure to adriamycin (16,17).

The occurrence of apoptosis in heart, kidney and small intestine was investigated in SHRs treated with doxorubicin (1 mg/kg/week for six, nine and 12 weeks) with and without pretreatment with the iron chelator (+)-1,2-bis(3,5-dioxopiperazinyl-1-yl)propane (dexrazoxane; ICRF-187) 25 mg/kg, intraperitoneally, given 30 min before doxorubicin. The results obtained by counting cells with positive nick-end labelling showed that, when given in cumulative doses of 9 and 12 (but not 6) mg/kg, doxorubicin induced significant toxicity in the heart, kidneys and intestine in association with apoptosis in epithelial cells of the intestinal mucosa and renal tubules but not in cardiac myocytes. At these doses nick end labelling in the heart was confined to occasional endothelial cells, interstitial dendritic cells and macrophages (18).

PROTECTION

Doxorubicin-induced overproduction of free radicals and subsequent damage to cellular membranes and the genome of cardiac myocytes can be prevented by a variety of agents. Probucol, a lipid-lowering agent and potent antioxidant, provides complete protection against adriamycin-induced cardiomyopathy in rats without interfering with the antitumour properties of this antibiotic. Melatonin protects against adriamycin-induced cardiomyopathy; it has been shown to affect zinc turnover (19). Zinc may act as an antioxidant, suggesting that myocardial zinc accumulation may be a protective response against adriamycin-induced oxidative stress (20). The peroxyl radical-scavenging activity of vitamin A was shown by pretreatment with 25 IU of vitamin A/kg, which significantly increased the survival rate of the animals (21). Carvedilol protects against doxorubicin-induced cardiomyopathy. This effect may be attributed to its antioxidant and lipid-lowering properties, not to its beta-blocking property (22). On the other hand, benidipine ameliorated doxorubicin-induced impairment of calcium dynamics, suggesting that benidipine, a long acting calcium antagonist, has potential clinical usefulness on doxorubicin-induced abnormal calcium handling (23).

The high susceptibility of cardiac muscle to anthracy-clines appears to be caused, at least in part, by the interaction of these drugs with intracellular iron. The suggestion that iron plays an important part in anthracycline cardiotoxicity has been strengthened by the observation that the chelator ICRF-187 has a potent cardioprotective effect. The role of iron in the cardiotoxicity of anthracyclines together with the possible role of iron chelation therapy as a cardioprotective strategy may also result in enhanced antitumour activity (24).