- Email: [email protected]
An Ovine Model of Toxic, Nonischemic Cardiomyopathy—Assessment by Cardiac Magnetic Resonance Imaging
Journal of Cardiac Failure Vol. 14 No. 9 2008
An Ovine Model of Toxic, Nonischemic CardiomyopathydAssessment by Cardiac Magnetic Resonance Imaging PETER J. PSALTIS, MBBS,1,2 ANGELO CARBONE, BSc,1 ADAM NELSON, BMedSc,1 DENNIS H. LAU, MBBS,1,2 JIM MANAVIS, BSc,3 JOHN FINNIE, BVSC, PhD,3 KAREN S. TEO, MBBS,1 LORRAINE MACKENZIE, PhD,1,2 PRASHANTHAN SANDERS, MBBS, PhD,1,2 STAN GRONTHOS, BSC, PhD,4 ANDREW C.W. ZANNETTINO, BSc, PhD,2,4 AND STEPHEN G. WORTHLEY, MBBS, PhD1,2 Adelaide, Australia
ABSTRACT Background: There is a paucity of published experience investigating novel treatment strategies in preclinical and clinical studies of nonischemic cardiomyopathy. We set out to validate an ovine model of doxorubicin-induced cardiomyopathy, using cardiac magnetic resonance (CMR) to assess cardiac function. Methods and Results: Ten Merino sheep (51 6 8 kg) underwent intracoronary infusions of doxorubicin (1 mg/kg dose) every 2 weeks. Cardiac magnetic resonance was performed at baseline and at 6 weeks after final doxorubicin dose, along with transthoracic echocardiography, measurement of right heart pressure, and cardiac output. After final CMR examination, heart specimens were harvested for histologic analysis. The total dose of doxorubicin administered per animal was 3.8 6 0.5 mg/kg. Two animals died prematurely during the study protocol, with evidence of myocarditis. In the remaining 8 sheep, left ventricular ejection fraction dropped from 46.2 6 4.7% to 31.3 6 8.5% (P ! .001), accompanied by reductions in fractional shortening (31.6 6 1.8% baseline versus 18.2 6 3.9% final, P ! .01), cardiac output (3.8 6 0.6 L/min versus 3.0 6 0.4 L/min, P ! .05) and right ventricular ejection fraction (39.5 6 5.6% versus 28.9 6 9.6%, P ! .05). However, significant end-diastolic dilatation of the left ventricle was not observed. Delayed gadolinium uptake was detected by CMR in 2 sheep, in a typical nonischemic pattern. Widespread, multifocal histologic abnormalities consisted of cardiomyocyte degeneration, vasculopathy, inflammatory infiltrates, and replacement fibrosis. Conclusions: Moderate-severe cardiac dysfunction was reproducibly achieved through high-dose intracoronary doxorubicin, with acceptable animal mortality. CMR provides a powerful tool for assessing myocardial function, structural remodeling, and viability in such models. (J Cardiac Fail 2008;14:785e795) Key Words: Cardiomyopathy, animal model, nonischemic, magnetic resonance imaging.
Cardiomyopathy and its clinical syndrome of congestive cardiac failure, remain major causes of patient morbidity and mortality, despite improvements in conventional management options, including pharmacotherapy, device therapy, and organ transplantation.1,2 Consequently, novel
treatment strategies, such as stem cell transplantation and gene therapy, continue to undergo intensive investigation as potential adjuvant treatments.3 Nonischemic cardiomyopathy contributes to approximately one third of cases of clinical heart failure and
From the 1Cardiovascular Research Centre, Royal Adelaide Hospital; Departments of Medicine and Physiology, The University of Adelaide; 3 Hanson Institute Centre for Neurological Diseases and 4Division of Hematology, Bone and Cancer Laboratories, Institute of Medical and Veterinary Science, Adelaide, Australia. Manuscript received May 4, 2008; revised manuscript received June 17, 2008; revised manuscript accepted June 30, 2008. Reprint requests: Dr Peter James Psaltis, Cardiovascular Research Centre, Department of Cardiology, Level 5, McEwin Building, Royal Adelaide Hospital, Adelaide, South Australia, Australia, 5000. Tel.: (þ61) 8 8222 5608; Fax: (þ61) 8 8222 2722; E-mail: [email protected] Dr Psaltis is supported by a Postgraduate Medical Scholarship from the National Health and Medical Research Council of Australia and the
National Heart Foundation of Australia and a Dawes Scholarship from the Royal Adelaide Hospital. Dr Lau is supported by a Postgraduate Medical Scholarship from the National Health and Medical Research Council of Australia, the Earl Bakken Electrophysiology Scholarship from the University of Adelaide and a Kidney Health Australia Biomedical Research Scholarship. Dr Mackenzie is supported by the Peter Doherty Fellowship, National Health and Medical Research Council of Australia. Dr Sanders is supported by the National Heart Foundation of Australia. 1071-9164/$ - see front matter Crown Copyright Ó 2008 Published by Elsevier Inc. All rights reserved. doi:10.1016/j.cardfail.2008.06.449
2
785
786 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 comprises various underlying etiologies, including idiopathic, genetic, infectious, metabolic, iatrogenic, and toxic processes.4 Despite its clinical burden, experimental investigation of cell-based therapy has been relatively understudied in nonischemic heart disease,5e7 necessitating further translational, preclinical research to complement basic scientific work and clinical trials. Appropriate large animal models of nonischemic cardiomyopathy that are stable, reproducible, and representative of the progressive nature of the clinical disease are therefore crucial. Traditional pacing-based strategies that induce ventricular dysfunction via tachycardia are limited by the rapidly reversible nature of the associated pathology.8 Alternative models have been attempted in a number of different animal species, including doxorubicin-induced, toxic cardiomyopathy. Doxorubicin’s cardiotoxic profile is well established both in the clinical setting and also from exhaustive in vitro studies.9,10 Although a previous study has described the use of intracoronary doxorubicin to induce heart failure in sheep, characterization was limited to transthoracic echocardiography and tissue-Doppler analysis.11 Cardiac magnetic resonance (CMR) imaging provides high-resolution assessment of the myocardium and is the current ‘‘gold standard’’ imaging modality for the noninvasive assessment of cardiac volumes and function12 and for documenting myocardial scar in both ischemic13 and nonischemic14 left ventricular dysfunction. Therefore it is an invaluable tool in preclinical research to ascertain if new models of heart failure are representative of cardiomyopathy in humans. In this study, we have undertaken detailed evaluation of a doxorubicin-induced cardiac injury model in sheep, with state of the art CMR and histopathologic analysis. Methods The study was approved by the Animal Ethics Committees of the Institute of Medical and Veterinary Services, Adelaide and the University of Adelaide, South Australia. Animal handling was carried out humanely in accordance with the ‘‘Principles of Laboratory Care’’ formulated by the National Society for Medical Research and the ‘‘Guide for the Care and Use of Laboratory Animals’’ prepared by the Institute of Laboratory Research and published by the National Institutes of Health. Experimental Protocol Ten merino sheep (weight range 40 to 63 kg) were included in the protocol. Doxorubicin (1 mg/kg) was infused via the coronary arterial route at fortnightly intervals, until animals were determined to have at least moderate left ventricular systolic dysfunction by transthoracic echocardiography (TTE). The primary measure of change in cardiac contractile function was obtained by CMR, which was performed both before the doxorubicin regimen and 6 weeks after the final dose had been administered. General Anesthesia Animals were fasted for 24 hours before general anesthesia, which was carried out for both investigations (echocardiography
and CMR) and invasive procedures (pericardial window, hemodynamic assessment, cardiac catheterization, and euthanasia). Anesthetic induction was achieved by intravenous administration of sodium thiopental (15e20 mg/kg), followed by endotracheal intubation and maintenance of anesthesia by inhalation of a mixture of isoflurane (2e3%) in 100% oxygen. Animals were mechanically ventilated with tidal volume of 10 mL/kg to maintain a partial pressure of expired carbon dioxide of approximately 40 mm Hg. Additional monitoring under anesthesia included pulse oximetry, continuous electrocardiography, and intra-arterial assessment of blood pressure. At completion, isoflurane was ceased, mechanical ventilation weaned, and extubation performed. Postoperative care after invasive surgeries included subcutaneous administration of ketoprofen and penicillin. Pericardial Window Before baseline CMR and commencement of doxorubicin dosing, sheep underwent creation of a small pericardial window to avoid inflammatory pericardial effusion (noted in previous studies).11 Left lateral thoracotomy was performed with the animal in the left lateral recumbent position. After making a linear incision in the 4th intercostal space, and dissecting through the fascial and muscular layers, the ribs were retracted and the pericardium visualized to enable excision of a 3 cm oval segment of pericardium using scissors. As the surgical wound was being closed in 3 layers, the lungs were mechanically ventilated by manual compression of ventilatory bag to ensure adequate expansion before final closure of the thoracic cavity. Animals recovered quickly without insertion of a chest tube. Antibiotics and analgesics were continued for 3 postoperative days. Doxorubicin Infusion The doxorubicin dosing strategy was established during pilot work. In these original studies, repeated administration of 0.75 mg/kg dose doxorubicin did not achieve the intended degree of cardiac dysfunction and so the higher dose regimen described as follows was employed. Venous blood specimens were taken immediately before each surgery to monitor cardiac enzymes, blood picture, electrolytes, and renal function. After this, peripheral arterial access with 5-Fr introducer sheath (Cordis Corp, Miami, FL) was obtained by percutaneous cannulation of the right femoral artery. A total of 100 IU/kg heparin was given by bolus into the arterial sheath and intra-arterial manometry commenced. An Amplatz diagnostic catheter (usually ALI) (Cordis Corp) was used to catheterize the left-sided coronary arteries under fluoroscopic guidance. Coronary artery engagement was confirmed by bolus administration of iodinated contrast media (Ultra-vist 360 Bayer Australia Ltd, New South Wales, Australia). To ensure stable delivery of the doxorubicin down the target vessels, the left anterior descending (‘‘left homonymous’’) and circumflex arteries were engaged selectively. Half of the medication was given to each artery in sequence. The total dose of doxorubicin (1 mg/kg in 50 mL 0.9% saline) was infused over 30 minutes. During dosing, catheter position was repeatedly visualized by fluoroscopy and records were taken of heart rate, intra-arterial blood pressure, and electrocardiogram appearance. When dosing had been completed and the catheter removed, hemostasis was achieved by manual compression and animals received antibiotic prophylaxis and analgesia, which were continued for 72 hours. Troponin-T and creatine kinase levels were checked 1 day after doxorubicin administration and clinical record sheets
Ovine Model of Toxic, Nonischemic Cardiomyopathy maintained throughout the protocol. Animals returned for 3 to 4 doses every 2 weeks, as guided by serial assessment of cardiac function by transthoracic echocardiography. CMR Imaging CMR scans were obtained using a Siemens Sonata 1.5 Tesla MR imaging system (Siemens Medical Solutions, Germany), allowing assessment of global and regional left and right ventricular ejection fractions, cardiac volumes (end-diastolic and endsystolic), and myocardial viability. Baseline imaging took place just before the first doxorubicin dose (after creation of the pericardial window) and follow-up scanning was performed 6 weeks after completion of dosing, just before euthanasia. Before CMR, sheep were anesthetized and had the wool over their left parasternal region shaved and the skin thoroughly cleansed with warm water, detergent, and ethanol to improve electrocardiogram lead adherence. They were then positioned in dorsal recumbency inside the scanner and mechanical ventilation commenced to ensure adequate breath-holds during image acquisition. The CMR protocols used have been described previously.12,13,15 The cine images were electrocardiogram-gated and consisted of steady-state free precessionebased sequences (TrueFISP sequences) with the following parameters: repetition time/echo time 52.05 ms/1.74 ms, flip angle 70 , matrix 256 150, 25 phases per cardiac cycle, field of view 380 mm with slice thickness 6 mm and interslice gap 4 mm through the ventricles. Heart rate was typically 90 to 100 beats/min (R-R0 interval 667 to 600 ms) during imaging and therefore the breath-hold times were between 8 and 12 seconds for all animals. A bolus of Gadolinium-DTPA (0.1 mmol/kg) was administered intravenously to acquire delayed contrast-enhanced images after a 10- to 15-minute delay, in the same views and positions as the cine images. The typical voxel size was 1.9 1.4 6 mm. The inversion delay was determined using an inversion recovery scout sequence, and was lengthened during scanning to maintain optimal nulling of normal myocardium, as described elsewhere.13,15 Therefore the inversion delays were between 260 and 320 ms. Standard and phase-sensitive imaging were performed, although the standard, delayed contrast-enhanced, inversion recovery gradient echo-based sequences were used for analyses. Ejection fractions and chamber volumes were analyzed by using Argus software (Leonardo workstation, Siemens Medical Solutions). The delayed contrast-enhanced images were analyzed visually and independently by two investigators with extensive CMR experience (KSLT, SGW). TTE TTE was performed at baseline and then used to monitor cardiac function and geometry on a serial basis during doxorubicin dosing, as well as at the end of the study (Acuson XP-128, 4 MHz probe, Siemens Medical Systems, Pennsylvania). Animals were fully anesthetized and mechanically ventilated for 30 minutes before imaging commenced with the animal placed in the right lateral decubitus position. Two-dimensional M-mode measurements were taken from right parasternal short axis views of the left ventricle just basal to the insertion of the papillary muscles, including left ventricular end-systolic and end-diastolic dimensions and left ventricular fractional shortening. For each parameter, measurements from 3 separate consecutive cardiac cycles were averaged.
Psaltis et al
787
Hemodynamic Measurements Right-sided cardiac pressure readings and assessment of cardiac output were performed immediately before commencement of doxorubicin dosing and at the end of the study, before euthanasia. The left external jugular vein was accessed by an 8-Fr introducer sheath with sidearm (Cordis Corp) and a 7.5-Fr thermodilution catheter (Edwards Lifesciences Corporation, Irvine, CA) was used to obtain right atrial, right ventricular, pulmonary artery, and pulmonary capillary wedge pressures as well as cardiac output. Cardiac outputs (Edwards 9520A Cardiac Output Monitor on loan from Professor G. Ludbrook, Department of Anesthesia, University of Adelaide, South Australia) were obtained at expired carbon dioxide pressure 40 mm Hg, from triplicate thermodilution measurements using iced 5% dextrose boluses (10 mL). Pathologic Examination Sheep were euthanized 6 weeks after their final intracoronary dose of doxorubicin by an intravenous overdose of sodium pentobarbital (60 mg/kg). Hearts were then immediately removed, perfusion-fixed with 4% paraformaldehyde containing 0.02% heparin, and immersed in a large volume of 10% buffered formalin. Transverse slices of the heart were collected at 1-cm intervals to ensure examination of a wide range of atrial and ventricular regions, and paraffin-embedded. Several 6-mm sections were cut and stained with hematoxylin and eosin. Duplicate sections were also stained by the Masson’s trichrome technique to demonstrate fibrosis and immunostained for extravasated albumin to detect edema, using goat anti-rat albumin (1:20,000) (Cappel) as the primary monoclonal antibody and biotinylated rabbit anti-goat immunoglobulin (1:500) (Dako, Denmark) as the secondary antibody; Appropriate positive and negative, isotype-matched, control monoclonal antibodies were used in this protocol. All slides were blind coded for independent examination by 2 pathologists. Statistical Analysis Parameters of cardiac geometry and function were recorded from baseline to the end of the study protocol, with the primary end point being the change in global left ventricular ejection fractions as determined by CMR. Data are presented as mean 6 SD. Comparisons between pre- and post-doxorubicin measurements were performed by dependent Student’s t-test (2-tailed), with P values of less than .05 considered statistically significant.
Results Ten animals underwent the intracoronary doxorubicin protocol. There were 2 deaths before study completion, both involving the sheep rapidly deteriorating overnight with few signs of heart failure in the preceding hours. In the first instance, the animal died 4 days after its second doxorubicin dose and the second nonsurvivor died under similar circumstances 4 days after its fourth dose. Necropsies from both animals revealed widespread inflammatory changes in the myocardium with congestion of the lungs and airways. Cause of death was concluded to be due to acute arrhythmia, possibly triggering acute pulmonary edema.
788 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 The 8 remaining sheep (mean starting weight 50.7 6 8.2 kg) received an average doxorubicin dose of 3.8 6 0.5 mg/ kg and were each followed for 6 weeks after their final doxorubicin dose. Systemic adverse effects from doxorubicin including wool loss (n 5 4) and self-limiting diarrhoea (n 5 2), were generally mild. Hematologic monitoring demonstrated stable hemoglobin (94.4 6 8.2 g/L pre-first dose versus 94.5 6 1.7 g/L post-final dose; P 5 NS) and platelet counts (201.0 6 54.0 109/L versus 215.0 6 31.2 109/L post-final dose; P 5 NS), but significant reduction in total white cell count 5.4 6 1.1 109/L versus 3.3 6 0.7 109/L; P 5 .02). Electrocardiographic Monitoring and Cardiac Enzymes
Electrocardiographic monitoring commonly demonstrated transient ST changes during intracoronary drug administration, although infusions were never associated with ectopy or arrhythmia. Heart rate, arterial blood pressure, and left ventricular systolic pressure remained consistent during dosing. Troponin-T values became elevated (O0.1 mg/L) after 53% of all doses (0% positive after first dose, 60% after second dose, 75% after third dose, 100% after fourth dose). Despite this, no hearts had evidence of myocardial infarction, either by CMR or histologic analysis. Late-onset sinus bradycardia was observed in 1 sheep and was persistent during its final CMR and euthanasia studies. Cardiac Function
High-quality electrocardiographic signals were able to be achieved in all animals during CMR imaging, although some positional changes of leads were required to minimize T-wave peaking within the scanner. Doxorubicin resulted in mean left ventricular ejection fractions reduction of 14.9% (range 6.0e28.7%) meeting our aim of inducing moderatesevere cardiac dysfunction (Table 1, Fig. 1). LV systolic dysfunction was global. Overt signs of heart failure were observed in 2 sheep during the later stages of doxorubicin dosing, necessitating frusemide. Right ventricular ejection Table 1. Summary of Results Baseline Weight 50.7 6 8.2 Heart rate beat3/min 103.7 6 10.6 PCWP (mm Hg) 7.8 6 2.3 Cardiac output(L/min) 3.8 6 0.6 Trans thoracic echocardiogram parameters LVEDD (mm) 43.4 6 5.1 LVESD (mm) 29.4 6 3.3 LVFS (%) 31.6 6 1.8 Cardiac magnetic resonance parameters LVEDV (ml) 68.5 6 14.0 LVESV (ml) 37.1 6 8.8 LVEF (%) 46.2 6 4.7 RVEF (%) 39.5 6 5.6
Final
Pvalue
6 6 6 6
8.7 4.5 1.9 0.4
NS NS NS !.05
43.8 6 2.4 35.8 6 2.9 18.2 6 3.9
NS !.005 !.01
6 6 6 6
NS NS !.001 !.05
49.6 105.1 8.0 3.0
69.3 47.7 31.3 28.9
15.8 13.2 8.5 9.6
Baseline and final results are presented as mean 6 standard deviation. LV, left ventricle; EDD, end-diastolic dimension; ESD, end-systolic dimension; FS, fractional shortening; EDV, end-diastolic volume; ESV, endsystolic volume; EF, ejection fraction; RV, right ventricle; NS, not significant.
fraction also significantly decreased by 10.6% from baseline (range 0e26.1%). Additional measures of doxorubicininduced systolic dysfunction included significant reduction in left ventricular fractional shortening and cardiac output. Hemodynamic measurements, including heart rate, arterial blood pressure, pulmonary capillary wedge pressure, and right-sided heart pressures did not significantly change during the study period. Although doxorubicin therapy resulted in increases in left ventricular end-systolic dimension, end-diastolic dilatation was not evident 6 weeks after completion of dosing. CMR assessment of myocardial viability did not show evidence of subendocardial or transmural, delayed enhancement of gadolinium in any sheep, indicating absence of myocardial infarction. However, 2 sheep had gadolinium enhancement in a mid-mural distribution, consistent with nonischemic injury (Fig. 2). Pathology
At necropsy, all treated hearts showed fibrinous adhesions to the pericardium and multifocal areas of pallor, both on epicardial surfaces and irregularly distributed on cut surfaces throughout the myocardium. Microscopically, in right and left ventricles, there was multifocal degeneration and necrosis of cardiac myocytes. Myocyte damage was frequently focal and well-circumscribed (Fig. 4C,D) but sometimes, especially in subendocardial sites, extensive (Fig. 3C). Affected myocytes were either markedly vacuolated (‘‘adria’’ cells) or showed coagulative necrosis characterized by swollen, granular or homogeneous, weakly eosinophilic cytoplasm, loss of striations, and nuclear hyperchromasia (Fig. 3A,B, Fig. 4F). There was substantial edema (confirmed by immunostaining for endogenous albumin) in selected foci of myocardial degeneration (Fig. 4A,B) and frequently a lymphoplasmacytic infiltrate (Fig. 3C, Fig. 4E). Anitchkov cells, with their characteristic ‘‘caterpillar’’ nuclei formed by a wavy ribbon of chromatin (Fig. 3D), were often found in areas of myocardial degeneration. Although the precise origin of these cells is uncertain, they probably represent either macrophages or an abortive attempt at myocyte regeneration. Notably in hearts that had received a fourth dose of doxorubicin, there was substantial replacement fibrosis of cardiomyocytes, well illustrated by Masson’s trichrome staining (Fig. 4C,D). A novel finding was frequent and severe, multivacuolation of Purkinje cells (Fig. 5C), sometimes attended by a lymphocytic infiltrate (Fig. 5D). Vascular changes, in arterioles and venules, were principally characterized by either a nonsuppurative vasculitis with medial myocyte degeneration, pyknosis of endothelial nuclei, and perivascular and mural lymphocytic infiltration (Fig. 5A), or severe, full-thickness degeneration of vessel walls (without an inflammatory reaction) and marked perivascular edema (Fig. 5B). In a few vessels, there was prominent proliferation of vasa vasorum. No thrombi were found in any cardiac blood vessels.
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
789
Fig. 1. (A) Example of steady-state free precession based cardiac magnetic resonance sequences (TrueFISP sequences). Images before and 6 weeks after completion of doxorubicin dosing. Top panel corresponds to pre-doxorubicin (left ventricular ejection fraction [LVEF] 48.5%, left ventricular end-diastolic volume [LVEDV] 53.9 mL, left ventricular end-systolic volume [LVESV] 27.8 mL). Long axis end-diastole (A) and end-systole (B) and short axis end-diastole (C) and end-systole (D). Bottom panel corresponds to post-doxorubicin (LVEF 32.1%, LVEDV 58.8 mL, LVESV 39.9 mL). Long axis end-diastole (E) and end-systole (F) and short axis end-diastole (G) and end-systole (H). LV, left ventricle; RV, right ventricle. (B) All sheep had reduction of LVEF 6 weeks after completion of doxorubicin dosing.
Discussion Animal Models of Cardiomyopathy
In this study, we describe an ovine model of anthracycline-induced cardiac dysfunction using a combination of CMR, hemodynamic, and histopathologic indices. Despite the development of numerous animal models of cardiomyopathy (CMP), none have quite fulfilled all of the criteria necessary to sufficiently mimic human heart failure. Animals used in preclinical research are typically young adults
that are not suffering from the various chronic health problems that often accompany heart failure in elderly human patients. In addition, many experimental protocols used to inflict damage to the heart do so abruptly giving rise to rapid changes in cardiac structure and function, rather than the slowly progressive remodeling that typifies the usual clinical syndrome of CHF. Nevertheless, animal studies are indispensable for complementing in vitro scientific research to provide crucial information relating to the pathophysiology and treatment of heart failure. Although the
790 Journal of Cardiac Failure Vol. 14 No. 9 November 2008
Fig. 2. Delayed myocardial enhancement of gadolinium on cardiac magnetic resonance images in a mid-mural distribution, consistent with nonischemic change. Long axis 4-chamber (A) and short axis mid-ventricular (B). Macroscopic appearance of the same heart specimen at necropsy showing mid-wall myocardial pallor (arrows) (C) and magnified view (D).
importance of small animal research is unquestionable, the successful development of new therapeutic strategies also requires evaluation and refinement in large animals that more closely resemble human anatomy, physiology, and function. Nonischemic, dilated cardiomyopathy (DCM) accounts for a substantial portion of patients with symptomatic cardiac dysfunction, and may arise from numerous etiologies.4,16 Its hallmark is an increase in left ventricular chamber radius-towall thickness ratio with resultant increases in left ventricular wall stress. As with its ischemic counterpart, there is a range of different approaches to inducing preclinical DCM in small and large animal species. Spontaneous models of disease are well recognized in the Syrian hamster17 and in large and giant breed dogs.18 Other well-described models have incorporated experimental tachycardia,19,20 hypertension,21 or mitral regurgitation.22 Chronic pacing tachycardia has emerged as a common method by which to induce DCM and has been applied to sheep, pigs, and dogs.20 Sustained
atrial or ventricular pacing at heart rates between 210 to 240 beats/min characteristically results in biventricular dilatation with reduced cardiac output.19,23 Characteristic pathophysiologic changes include contractile dysfunction of isolated myocytes, abnormal intracellular calcium regulation, abnormalities in b-adrenergic responsiveness, and depletion of high-energy phosphates. Despite advantages, which include a progressive and predictable degree of left ventricular dilation and pump dysfunction and neurohormonal activation, rapid pacing models do not manifest the full spectrum of CHF. Structural and histologic changes can be quite divergent from those observed in clinical CMP and can partly recover on cessation or interruption of pacing.24 In addition, pacing models are not ideal for analysis by CMR, which is now considered a ‘‘gold standard’’ in myocardial assessment for both preclinical and clinical studies. CMR provides superior tissue resolution compared to other noninvasive imaging modalities and can be used to obtain multidimensional information during single studies.12
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
791
Fig. 3. Sheep; ventricle. Large focus of myocardial degeneration (A) with higher power (B) showing severely vacuolated myocytes (black arrow) and myocytes undergoing coagulative necrosis (white arrow). Extensive area of subendocardial degeneration and fibroplasia (C) with lymphoplasmacytic infiltration. Anichkov cells (arrow) with ‘‘caterpillar’’ nuclei (D). (Hematoxylin and eosin).
Doxorubicin-Induced Cardiomyopathy
Toxic cardiomyopathy is an alternative model of nonischemic heart failure that does not have some of the limitations inherent in the pacing-tachycardia strategy. Doxorubicin (Adriamycin) has traditionally been the ‘‘toxin of choice’’ across different animal species. This quinonecontaining anthracycline antibiotic is most frequently used as a chemotherapeutic agent for various types of solid tumors and hematologic malignancies. Its major clinical shortcoming is dose-related cardiotoxicity, which can manifest both as acute myocarditis and progressive, chronic CMP. In human patients, cardiotoxicity is largely determined by cumulative dose exposure, such that life-threatening CHF effects 10% of patients who have received more than 550 mg/m2 systemic dose.25 The mechanistic pathways mediating such toxicity are multiple and complex, with damage to mitochondrial respiratory chain function and the liberation of highly reactive, free oxygen radicals playing a key role in initiating a cascade of deleterious subcellular sequelae. Cardiomyocyte morphological changes include cytoplasmic vacuolation and myofibrillar loss, with the end-result being apoptosis of both myocytes and other cardiac-relevant cells, such as endothelial cells and replacement fibrosis.9,26 Although several groups have reported successful induction of cardiac dysfunction in animals administered this drug, their strategies have differed considerably especially
with respect to dosing regimen and route of drug administration. In small animal models of myocarditis and heart failure, doxorubicin has usually been administered by either single intraperitoneal dose5 or serial intravenous doses.7,27 Repeat intravenous administration has also been used in dogs,28 goats, and sheep, sometimes in combination with arteriovenous anastomosis29,30; however, the high doses of medication required have been associated with substantial systemic effects and mortality. The coronary route of administration therefore has been applied to enable lower doses to be given, with high firstpass concentrations in the myocardium and fewer systemic effects.11,31e33 In the early canine models, dosing was typically weekly for up to 5 doses and oral verapamil therapy was sometimes combined with doxorubicin exposure to augment cardiotoxicity.32,34 Mortality was still substantial (20e60%), often because of fatal arrhythmia. The first description of doxorubicin-induced cardiomyopathy by coronary administration in sheep was reported recently by Borenstein et al,11 who partly attributed their improvement in mortality rate to the creation of a pericardial window, which was designed to prevent doxorubicin-induced inflammatory pericardial effusion and tamponade. Assessment of cardiac dysfunction in this study was both by echocardiography, including tissue Doppler imaging, and left ventriculography. Animals received a mean of 2.5 doses of 0.75
792 Journal of Cardiac Failure Vol. 14 No. 9 November 2008
Fig. 4. Sheep; ventricle. Focal myocardial degeneration and cardiomyocyte loss (A) with edematous area immunostained for endogenous albumin (B) showing marked albumin extravasation. Well-circumscribed focus of myocardial damage (C) with substantial reparative fibrosis evident by Masson’s trichrome staining (D). Focal myocardial injury showing lymphoplasmacytic infiltrate and residual degenerate myocytes (E) and same area stained with Masson’s trichrome showing effete myocytes marooned in collagenous connective tissue (F).
mg/kg, resulting in a mean ejection fraction decline of approximately 16%. In our pilot work, using Merino sheep, initial dosing with 0.75 mg/kg dose did not achieve the intended deterioration in left ventricular contractility. Consequently, doses were increased to 1 mg/kg and sheep required more doses than was the case in the Borenstein description of this model. The nonsurvival rate we observed (20%) was favorable compared with the range reported by earlier studies, with both deaths in our group likely to be due to arrhythmia, accompanied by acute pulmonary edema. Surviving sheep demonstrated relatively mild systemic side effects and weight was well maintained throughout the full study protocol.
Unlike swine, the coronary anatomy of sheep is not as similar to humans, with sheep demonstrating a left dominant circulation with large circumflex artery. In our experience, the left main coronary artery in Merino sheep is typically short and difficult to maintain stable catheter engagement over a prolonged period. Consequently, we selectively engaged and dosed the left anterior descending and left circumflex arteries in sequence, using an AL I catheter. Although high-dose doxorubicin (1 mg/kg) was administered over a shorter period than in other studies, we did not observe intraoperative arrhythmia and did not lose any sheep directly from intraprocedural complications. Heart rate, blood pressure, and left ventricular systolic
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
793
Fig. 5. Sheep; ventricle. Nonsuppurative vasculitis (A) with lymphocytic infiltration of degenerate wall (arrow) and severe vasculopathy (B) showing dissolution of the vessel wall and marked perivascular edema. Severe vacuolation (arrow) of Purkinje cells (C) with higher power view as inset and a lymphocytic inflammatory reaction (D). (Hematoxylin and eosin).
pressure did not fluctuate substantially during individual doses and, although ST depression was occasionally observed, this phenomenon was transient and almost never residual at the end of surgery. Evaluation by CMR
This is the first description of the utility of CMR in a preclinical model of doxorubicin-induced CMP. Although we used serial TTE to monitor the progressive nature of cardiac decline and to indicate which sheep might require a fourth dose of doxorubicin, echocardiographic image quality was variable, such that right parasternal images were not always satisfactorily obtainable. Magnetic resonance imaging provided superior resolution than TTE for the evaluation of cardiac geometry and function. It also provided a noninvasive measure of viability, to complement histopathology in ruling out transmural myocardial infarction, through the assessment of delayed gadolinium enhancement. Two of 8 sheep displayed patchy, mid-mural uptake of gadolinium, which closely mirrors the pattern and prevalence of late gadolinium uptake reported in human patients with DCM.35,36 The exclusion of myocardial infarction was especially important given the frequency of elevated troponin-T levels after doxorubicin doses and led us to infer that this was probably the result of progressive doxorubicin-induced myocarditis, rather than infarction.
Although there was some individual variation in response to high-dose intracoronary doxorubicin, all sheep had global deterioration in left ventricular function. In addition, we also observed reduction in mean right ventricular ejection fraction, not reported by other studies in which doxorubicin has been administered without adjuvant arteriovenous fistula creation.30 Notably, we did not achieve end-diastolic dilation of the left ventricle, contrasting the previous report by Borenstein et al, who adopted a lower dose protocol, with fewer doses and shorter follow-up, but obtained their measurements by 2-dimensional echocardiography.11 In future studies, we intend to follow animals for longer periods to determine whether dilation becomes a later feature of our model. The pathologic findings observed were consistent with those of previous studies and with the changes reported in human patients suffering anthracycline-induced cardiomyopathy. Lesions were characterized by either multifocal degeneration and necrosis of cardiomyocytes with attendant inflammation or myocyte loss and replacement fibrosis, and nonocclusive vasculopathy. It has been previously demonstrated that these doxorubicin-induced effects are largely irreversible and progressive over time, resulting in a more suitable pathologic substrate for preclinical trials of heart failure therapy (eg, antifibrotic drugs and cellular cardiomyoplasty) than traditional pacing models. Another notable
794 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 finding of our study was the consistent damage caused by doxorubicin to Purkinje fibres. This has not been described in previous in vivo reports, although anthracyclines are known to affect Purkinje fibre action potentials in vitro.37 Such changes may contribute to the substantial rate of arrhythmia known to complicate anthracycline cardiomyopathy and will be the subject of future electrophysiologic testing in our model.
6.
7.
8.
Study Limitations
A limitation of our description of doxorubicin cardiotoxicity is the lack of invasive left ventricular hemodynamic analysis by conductance catheters and the absence of assessment of myocardial energo-mechanics. The sequential intracoronary dosing regime requires multiple repeat anesthetics and measurements performed in the anesthetized state may differ from those at conscious and closechest state. Although we were satisfied by the progressive response of cardiac systolic function to doxorubicin, there inevitably was some variability in the degree of response between individual sheep and this supports the need to titrate the number of doses to obtain the desired effect.
9.
10.
11.
12.
13.
Conclusion In conclusion, we describe an ovine model of moderatesevere biventricular dysfunction, induced by repeat coronary administration of doxorubicin. Moreover, we demonstrate the preclinical utility of CMR for noninvasive and comprehensive assessment of cardiac anatomy, function, and viability. This information, coupled by histopathologic analysis, has shown this model to be representative of nonischemic cardiomyopathy seen in humans.
14.
15.
16. 17.
Acknowledgment The authors would also like to thank Dr Tim Kuchel, Ms Jodie Dier, Ms Melissa Gourlay, and Mr Adrian Hines (Veterinary Services Division, IMVS); Ms Kerry Williams and Mr Leigh Penney (Perrett Medical Imaging, Wakefield Street, Adelaide); and Dr Zhao Cai (Hanson Institute Centre for Neurological Diseases, IMVS) for their assistance during this study.
18.
19.
20. 21.
References 22. 1. Fonarow GC. Heart failure: recent advances in prevention and treatment. Rev Cardiovasc Med 2000;1:25e33, 54. 2. Jessup M., Brozena S. Heart failure. N Engl J.M. 2003;348: 2007e2018. 3. Laflamme MA, Murry CE. Regenerating the heart. Nat Biotechnol 2005;23:845e56. 4. Follath F. Ischemic versus non-ischemic heart failure: should the etiology be determined? Heart Fail Mon 2001;1:122e5. 5. Agbulut O, Menot ML, Li Z, Marotte F, Paulin D, Hagege AA, et al. Temporal patterns of bone marrow cell differentiation following
23.
24.
transplantation in doxorubicin-induced cardiomyopathy. Cardiovasc Res 2003;58:451e9. Borenstein N, Chetboul V, Bruneval P, Hekmati M, Tissier R, Behr L, et al. Non-cultured cell transplantation in an ovine model of nonischemic heart failure. Eur J Cardiothorac Surg 2007;31:444e51. Aupperle H, Garbade J, Schubert A, Barten M, Dhein S, Schoon HA, et al. Effects of autologous stem cells on immunohistochemical patterns and gene expression of metalloproteinases and their tissue inhibitors in doxorubicin cardiomyopathy in a rabbit model. Vet Pathol 2007;44:494e503. Yarbrough WM, Spinale FG. Large animal models of congestive heart failure: a critical step in translating basic observations into clinical applications. J Nucl Cardiol 2003;10:77e86. Takemura G, Fujiwara H. Doxorubicin-induced cardiomyopathy from the cardiotoxic mechanisms to management. Progress Cardiovasc Dis 2007;49:330e52. Swain SM, Whaley FS, Ewer MS. Congestive heart failure in patients treated with doxorubicin: a retrospective analysis of three trials. Cancer 2003;97:2869e79. Borenstein N, Bruneval P, Behr L, Laborde F, Montarras D, Daures JP, et al. An ovine model of chronic heart failure: echocardiographic and tissue Doppler imaging characterization. J Cardiac Surg 2006;21: 50e6. Teo KS, Carbone A, Piantadosi C, Chew DP, Hammett CJ, Brown MA, et al. Cardiac MRI assessment of left and right ventricular parameters in healthy Australian normal volunteers. Heart Lung Circ 2008;17:313e7. Kim RJ, Wu E, Rafael A, Chen EL, Parker MA, Simonetti O, et al. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med 2000;343: 1445e53. McCrohon JA, Moon JC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJ, et al. Differentiation of heart failure related to dilated cardiomyopathy and coronary artery disease using gadoliniumenhanced cardiovascular magnetic resonance. Circulation 2003; 108:54e9. Jerosch-Herold M, Seethamraju RT, Swingen CM, Wilke NM, Stillman AE. Analysis of myocardial perfusion MRI. J Magn Reson Imaging 2004;19:758e70. Eriksson H. Heart failure: a growing public health problem. J Intern Med 1995;237:135e41. Gertz EW, Stam AC Jr, Sonnenblick EH. A quantitative and qualitative defect in the sarcoplasmic reticulum in the hereditary cardiomyopathy of the Syrian hamster. Biochem Biophys Res Communi 1970;40:746e53. Tidholm A, Svensson H, Sylven C. Survival and prognostic factors in 189 dogs with dilated cardiomyopathy. J Am Animal Hosp Assoc 1997;33:364e8. Yamamoto K, Burnett JC Jr, Meyer LM, Sinclair L, Stevens TL, Redfield MM. Ventricular remodeling during development and recovery from modified tachycardia-induced cardiomyopathy model. Am J Physiol 1996;271:R1529e34. Recchia FA, Lionetti V. Animal models of dilated cardiomyopathy for translational research. Vet Res Commun 2007;31(Suppl 1):35e41. Aoyagi T, Mirsky I, Flanagan MF, Currier JJ, Colan SD, Fujii AM. Myocardial function in immature and mature sheep with pressureoverload hypertrophy. Am J Physiol 1992;262:H1036e48. Nakano K, Swindle MM, Spinale F, Ishihara K, Kanazawa S, Smith A, et al. Depressed contractile function due to canine mitral regurgitation improves after correction of the volume overload. J Clin Invest 1991; 87:2077e86. Tanaka R, Spinale FG, Crawford FA, Zile MR. Effect of chronic supraventricular tachycardia on left ventricular function and structure in newborn pigs. J Am Coll Cardiol 1992;20:1650e60. Spinale FG, Holzgrefe HH, Mukherjee R, Arthur SR, Child MJ, Powell JR, et al. LV and myocyte structure and function after early recovery from tachycardia-induced cardiomyopathy. Am J Physiol 1995; 268:H836e47.
Ovine Model of Toxic, Nonischemic Cardiomyopathy 25. Singal PK, Iliskovic N. Doxorubicin-induced cardiomyopathy. N Engl J Med 1998;339:900e5. 26. Kalyanaraman B, Joseph J, Kalivendi S, Wang S, Konorev E, Kotamraju S. Doxorubicin-induced apoptosis: implications in cardiotoxicity. Mol Cell Biochem 2002;234-235:119e24. 27. Schwarz ER, Pollick C, Dow J, Patterson M, Birnbaum Y, Kloner RA. A small animal model of non-ischemic cardiomyopathy and its evaluation by transthoracic echocardiography. Cardiovasc Res 1998;39: 216e23. 28. Ditchey RV, LeWinter MM, Higgins CB. Acute effects of doxorubicin (Adriamycin) on left ventricular function in dogs. Int J Cardiol 1984;6: 341e53. 29. Tessier D, Lajos P, Braunberger E, Pouchelon JL, Carpentier A, Chachques JC, et al. Induction of chronic cardiac insufficiency by arteriovenous fistula and doxorubicin administration. J Cardiac Surg 2003;18:307e11. 30. Chekanov VS. A stable model of chronic bilateral ventricular insufficiency (dilated cardiomyopathy) induced by arteriovenous anastomosis and doxorubicin administration in sheep. J Thorac Cardiovasc Surg 1999;117:198e9.
Psaltis et al
795
31. Toyoda Y, Okada M, Kashem MA. A canine model of dilated cardiomyopathy induced by repetitive intracoronary doxorubicin administration. J Thorac Cardiovasc Surg 1998;115:1367e73. 32. Monnet E, Orton EC. A canine model of heart failure by intracoronary Adriamycin injection: hemodynamic and energetic results. J Cardiac Fail 1999;5:255e64. 33. Magovern JA, Christlieb IY, Badylak SF, Lantz GC, Kao RL. A model of left ventricular dysfunction caused by intracoronary Adriamycin. Ann Thorac Surg 1992;53:861e3. 34. Bright JM, Buss DD. Effects of verapamil on chronic doxorubicininduced cardiotoxicity in dogs. J Natl Cancer Inst 1990;82:963e4. 35. Stork A, Mullerleile K, Bansmann PM, Koops A, Meinertz T, Adam G, et al. [Patterns of delayed-enhancement in MRI of ischemic and non-ischemic cardiomyopathies]. Rofo 2007;179:21e30. 36. Gottlieb I, Macedo R, Bluemke DA, Lima JA. Magnetic resonance imaging in the evaluation of non-ischemic cardiomyopathies: current applications and future perspectives. Heart Fail Rev 2006;11:313e23. 37. Rabkin SW, Bose D. Adriamycin induced alterations in canine Purkinje fiber action potential. Res Commun Chem Pathol Pharmacol 1981;34:55e67.
An Ovine Model of Toxic, Nonischemic CardiomyopathydAssessment by Cardiac Magnetic Resonance Imaging PETER J. PSALTIS, MBBS,1,2 ANGELO CARBONE, BSc,1 ADAM NELSON, BMedSc,1 DENNIS H. LAU, MBBS,1,2 JIM MANAVIS, BSc,3 JOHN FINNIE, BVSC, PhD,3 KAREN S. TEO, MBBS,1 LORRAINE MACKENZIE, PhD,1,2 PRASHANTHAN SANDERS, MBBS, PhD,1,2 STAN GRONTHOS, BSC, PhD,4 ANDREW C.W. ZANNETTINO, BSc, PhD,2,4 AND STEPHEN G. WORTHLEY, MBBS, PhD1,2 Adelaide, Australia
ABSTRACT Background: There is a paucity of published experience investigating novel treatment strategies in preclinical and clinical studies of nonischemic cardiomyopathy. We set out to validate an ovine model of doxorubicin-induced cardiomyopathy, using cardiac magnetic resonance (CMR) to assess cardiac function. Methods and Results: Ten Merino sheep (51 6 8 kg) underwent intracoronary infusions of doxorubicin (1 mg/kg dose) every 2 weeks. Cardiac magnetic resonance was performed at baseline and at 6 weeks after final doxorubicin dose, along with transthoracic echocardiography, measurement of right heart pressure, and cardiac output. After final CMR examination, heart specimens were harvested for histologic analysis. The total dose of doxorubicin administered per animal was 3.8 6 0.5 mg/kg. Two animals died prematurely during the study protocol, with evidence of myocarditis. In the remaining 8 sheep, left ventricular ejection fraction dropped from 46.2 6 4.7% to 31.3 6 8.5% (P ! .001), accompanied by reductions in fractional shortening (31.6 6 1.8% baseline versus 18.2 6 3.9% final, P ! .01), cardiac output (3.8 6 0.6 L/min versus 3.0 6 0.4 L/min, P ! .05) and right ventricular ejection fraction (39.5 6 5.6% versus 28.9 6 9.6%, P ! .05). However, significant end-diastolic dilatation of the left ventricle was not observed. Delayed gadolinium uptake was detected by CMR in 2 sheep, in a typical nonischemic pattern. Widespread, multifocal histologic abnormalities consisted of cardiomyocyte degeneration, vasculopathy, inflammatory infiltrates, and replacement fibrosis. Conclusions: Moderate-severe cardiac dysfunction was reproducibly achieved through high-dose intracoronary doxorubicin, with acceptable animal mortality. CMR provides a powerful tool for assessing myocardial function, structural remodeling, and viability in such models. (J Cardiac Fail 2008;14:785e795) Key Words: Cardiomyopathy, animal model, nonischemic, magnetic resonance imaging.
Cardiomyopathy and its clinical syndrome of congestive cardiac failure, remain major causes of patient morbidity and mortality, despite improvements in conventional management options, including pharmacotherapy, device therapy, and organ transplantation.1,2 Consequently, novel
treatment strategies, such as stem cell transplantation and gene therapy, continue to undergo intensive investigation as potential adjuvant treatments.3 Nonischemic cardiomyopathy contributes to approximately one third of cases of clinical heart failure and
From the 1Cardiovascular Research Centre, Royal Adelaide Hospital; Departments of Medicine and Physiology, The University of Adelaide; 3 Hanson Institute Centre for Neurological Diseases and 4Division of Hematology, Bone and Cancer Laboratories, Institute of Medical and Veterinary Science, Adelaide, Australia. Manuscript received May 4, 2008; revised manuscript received June 17, 2008; revised manuscript accepted June 30, 2008. Reprint requests: Dr Peter James Psaltis, Cardiovascular Research Centre, Department of Cardiology, Level 5, McEwin Building, Royal Adelaide Hospital, Adelaide, South Australia, Australia, 5000. Tel.: (þ61) 8 8222 5608; Fax: (þ61) 8 8222 2722; E-mail: [email protected] Dr Psaltis is supported by a Postgraduate Medical Scholarship from the National Health and Medical Research Council of Australia and the
National Heart Foundation of Australia and a Dawes Scholarship from the Royal Adelaide Hospital. Dr Lau is supported by a Postgraduate Medical Scholarship from the National Health and Medical Research Council of Australia, the Earl Bakken Electrophysiology Scholarship from the University of Adelaide and a Kidney Health Australia Biomedical Research Scholarship. Dr Mackenzie is supported by the Peter Doherty Fellowship, National Health and Medical Research Council of Australia. Dr Sanders is supported by the National Heart Foundation of Australia. 1071-9164/$ - see front matter Crown Copyright Ó 2008 Published by Elsevier Inc. All rights reserved. doi:10.1016/j.cardfail.2008.06.449
2
785
786 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 comprises various underlying etiologies, including idiopathic, genetic, infectious, metabolic, iatrogenic, and toxic processes.4 Despite its clinical burden, experimental investigation of cell-based therapy has been relatively understudied in nonischemic heart disease,5e7 necessitating further translational, preclinical research to complement basic scientific work and clinical trials. Appropriate large animal models of nonischemic cardiomyopathy that are stable, reproducible, and representative of the progressive nature of the clinical disease are therefore crucial. Traditional pacing-based strategies that induce ventricular dysfunction via tachycardia are limited by the rapidly reversible nature of the associated pathology.8 Alternative models have been attempted in a number of different animal species, including doxorubicin-induced, toxic cardiomyopathy. Doxorubicin’s cardiotoxic profile is well established both in the clinical setting and also from exhaustive in vitro studies.9,10 Although a previous study has described the use of intracoronary doxorubicin to induce heart failure in sheep, characterization was limited to transthoracic echocardiography and tissue-Doppler analysis.11 Cardiac magnetic resonance (CMR) imaging provides high-resolution assessment of the myocardium and is the current ‘‘gold standard’’ imaging modality for the noninvasive assessment of cardiac volumes and function12 and for documenting myocardial scar in both ischemic13 and nonischemic14 left ventricular dysfunction. Therefore it is an invaluable tool in preclinical research to ascertain if new models of heart failure are representative of cardiomyopathy in humans. In this study, we have undertaken detailed evaluation of a doxorubicin-induced cardiac injury model in sheep, with state of the art CMR and histopathologic analysis. Methods The study was approved by the Animal Ethics Committees of the Institute of Medical and Veterinary Services, Adelaide and the University of Adelaide, South Australia. Animal handling was carried out humanely in accordance with the ‘‘Principles of Laboratory Care’’ formulated by the National Society for Medical Research and the ‘‘Guide for the Care and Use of Laboratory Animals’’ prepared by the Institute of Laboratory Research and published by the National Institutes of Health. Experimental Protocol Ten merino sheep (weight range 40 to 63 kg) were included in the protocol. Doxorubicin (1 mg/kg) was infused via the coronary arterial route at fortnightly intervals, until animals were determined to have at least moderate left ventricular systolic dysfunction by transthoracic echocardiography (TTE). The primary measure of change in cardiac contractile function was obtained by CMR, which was performed both before the doxorubicin regimen and 6 weeks after the final dose had been administered. General Anesthesia Animals were fasted for 24 hours before general anesthesia, which was carried out for both investigations (echocardiography
and CMR) and invasive procedures (pericardial window, hemodynamic assessment, cardiac catheterization, and euthanasia). Anesthetic induction was achieved by intravenous administration of sodium thiopental (15e20 mg/kg), followed by endotracheal intubation and maintenance of anesthesia by inhalation of a mixture of isoflurane (2e3%) in 100% oxygen. Animals were mechanically ventilated with tidal volume of 10 mL/kg to maintain a partial pressure of expired carbon dioxide of approximately 40 mm Hg. Additional monitoring under anesthesia included pulse oximetry, continuous electrocardiography, and intra-arterial assessment of blood pressure. At completion, isoflurane was ceased, mechanical ventilation weaned, and extubation performed. Postoperative care after invasive surgeries included subcutaneous administration of ketoprofen and penicillin. Pericardial Window Before baseline CMR and commencement of doxorubicin dosing, sheep underwent creation of a small pericardial window to avoid inflammatory pericardial effusion (noted in previous studies).11 Left lateral thoracotomy was performed with the animal in the left lateral recumbent position. After making a linear incision in the 4th intercostal space, and dissecting through the fascial and muscular layers, the ribs were retracted and the pericardium visualized to enable excision of a 3 cm oval segment of pericardium using scissors. As the surgical wound was being closed in 3 layers, the lungs were mechanically ventilated by manual compression of ventilatory bag to ensure adequate expansion before final closure of the thoracic cavity. Animals recovered quickly without insertion of a chest tube. Antibiotics and analgesics were continued for 3 postoperative days. Doxorubicin Infusion The doxorubicin dosing strategy was established during pilot work. In these original studies, repeated administration of 0.75 mg/kg dose doxorubicin did not achieve the intended degree of cardiac dysfunction and so the higher dose regimen described as follows was employed. Venous blood specimens were taken immediately before each surgery to monitor cardiac enzymes, blood picture, electrolytes, and renal function. After this, peripheral arterial access with 5-Fr introducer sheath (Cordis Corp, Miami, FL) was obtained by percutaneous cannulation of the right femoral artery. A total of 100 IU/kg heparin was given by bolus into the arterial sheath and intra-arterial manometry commenced. An Amplatz diagnostic catheter (usually ALI) (Cordis Corp) was used to catheterize the left-sided coronary arteries under fluoroscopic guidance. Coronary artery engagement was confirmed by bolus administration of iodinated contrast media (Ultra-vist 360 Bayer Australia Ltd, New South Wales, Australia). To ensure stable delivery of the doxorubicin down the target vessels, the left anterior descending (‘‘left homonymous’’) and circumflex arteries were engaged selectively. Half of the medication was given to each artery in sequence. The total dose of doxorubicin (1 mg/kg in 50 mL 0.9% saline) was infused over 30 minutes. During dosing, catheter position was repeatedly visualized by fluoroscopy and records were taken of heart rate, intra-arterial blood pressure, and electrocardiogram appearance. When dosing had been completed and the catheter removed, hemostasis was achieved by manual compression and animals received antibiotic prophylaxis and analgesia, which were continued for 72 hours. Troponin-T and creatine kinase levels were checked 1 day after doxorubicin administration and clinical record sheets
Ovine Model of Toxic, Nonischemic Cardiomyopathy maintained throughout the protocol. Animals returned for 3 to 4 doses every 2 weeks, as guided by serial assessment of cardiac function by transthoracic echocardiography. CMR Imaging CMR scans were obtained using a Siemens Sonata 1.5 Tesla MR imaging system (Siemens Medical Solutions, Germany), allowing assessment of global and regional left and right ventricular ejection fractions, cardiac volumes (end-diastolic and endsystolic), and myocardial viability. Baseline imaging took place just before the first doxorubicin dose (after creation of the pericardial window) and follow-up scanning was performed 6 weeks after completion of dosing, just before euthanasia. Before CMR, sheep were anesthetized and had the wool over their left parasternal region shaved and the skin thoroughly cleansed with warm water, detergent, and ethanol to improve electrocardiogram lead adherence. They were then positioned in dorsal recumbency inside the scanner and mechanical ventilation commenced to ensure adequate breath-holds during image acquisition. The CMR protocols used have been described previously.12,13,15 The cine images were electrocardiogram-gated and consisted of steady-state free precessionebased sequences (TrueFISP sequences) with the following parameters: repetition time/echo time 52.05 ms/1.74 ms, flip angle 70 , matrix 256 150, 25 phases per cardiac cycle, field of view 380 mm with slice thickness 6 mm and interslice gap 4 mm through the ventricles. Heart rate was typically 90 to 100 beats/min (R-R0 interval 667 to 600 ms) during imaging and therefore the breath-hold times were between 8 and 12 seconds for all animals. A bolus of Gadolinium-DTPA (0.1 mmol/kg) was administered intravenously to acquire delayed contrast-enhanced images after a 10- to 15-minute delay, in the same views and positions as the cine images. The typical voxel size was 1.9 1.4 6 mm. The inversion delay was determined using an inversion recovery scout sequence, and was lengthened during scanning to maintain optimal nulling of normal myocardium, as described elsewhere.13,15 Therefore the inversion delays were between 260 and 320 ms. Standard and phase-sensitive imaging were performed, although the standard, delayed contrast-enhanced, inversion recovery gradient echo-based sequences were used for analyses. Ejection fractions and chamber volumes were analyzed by using Argus software (Leonardo workstation, Siemens Medical Solutions). The delayed contrast-enhanced images were analyzed visually and independently by two investigators with extensive CMR experience (KSLT, SGW). TTE TTE was performed at baseline and then used to monitor cardiac function and geometry on a serial basis during doxorubicin dosing, as well as at the end of the study (Acuson XP-128, 4 MHz probe, Siemens Medical Systems, Pennsylvania). Animals were fully anesthetized and mechanically ventilated for 30 minutes before imaging commenced with the animal placed in the right lateral decubitus position. Two-dimensional M-mode measurements were taken from right parasternal short axis views of the left ventricle just basal to the insertion of the papillary muscles, including left ventricular end-systolic and end-diastolic dimensions and left ventricular fractional shortening. For each parameter, measurements from 3 separate consecutive cardiac cycles were averaged.
Psaltis et al
787
Hemodynamic Measurements Right-sided cardiac pressure readings and assessment of cardiac output were performed immediately before commencement of doxorubicin dosing and at the end of the study, before euthanasia. The left external jugular vein was accessed by an 8-Fr introducer sheath with sidearm (Cordis Corp) and a 7.5-Fr thermodilution catheter (Edwards Lifesciences Corporation, Irvine, CA) was used to obtain right atrial, right ventricular, pulmonary artery, and pulmonary capillary wedge pressures as well as cardiac output. Cardiac outputs (Edwards 9520A Cardiac Output Monitor on loan from Professor G. Ludbrook, Department of Anesthesia, University of Adelaide, South Australia) were obtained at expired carbon dioxide pressure 40 mm Hg, from triplicate thermodilution measurements using iced 5% dextrose boluses (10 mL). Pathologic Examination Sheep were euthanized 6 weeks after their final intracoronary dose of doxorubicin by an intravenous overdose of sodium pentobarbital (60 mg/kg). Hearts were then immediately removed, perfusion-fixed with 4% paraformaldehyde containing 0.02% heparin, and immersed in a large volume of 10% buffered formalin. Transverse slices of the heart were collected at 1-cm intervals to ensure examination of a wide range of atrial and ventricular regions, and paraffin-embedded. Several 6-mm sections were cut and stained with hematoxylin and eosin. Duplicate sections were also stained by the Masson’s trichrome technique to demonstrate fibrosis and immunostained for extravasated albumin to detect edema, using goat anti-rat albumin (1:20,000) (Cappel) as the primary monoclonal antibody and biotinylated rabbit anti-goat immunoglobulin (1:500) (Dako, Denmark) as the secondary antibody; Appropriate positive and negative, isotype-matched, control monoclonal antibodies were used in this protocol. All slides were blind coded for independent examination by 2 pathologists. Statistical Analysis Parameters of cardiac geometry and function were recorded from baseline to the end of the study protocol, with the primary end point being the change in global left ventricular ejection fractions as determined by CMR. Data are presented as mean 6 SD. Comparisons between pre- and post-doxorubicin measurements were performed by dependent Student’s t-test (2-tailed), with P values of less than .05 considered statistically significant.
Results Ten animals underwent the intracoronary doxorubicin protocol. There were 2 deaths before study completion, both involving the sheep rapidly deteriorating overnight with few signs of heart failure in the preceding hours. In the first instance, the animal died 4 days after its second doxorubicin dose and the second nonsurvivor died under similar circumstances 4 days after its fourth dose. Necropsies from both animals revealed widespread inflammatory changes in the myocardium with congestion of the lungs and airways. Cause of death was concluded to be due to acute arrhythmia, possibly triggering acute pulmonary edema.
788 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 The 8 remaining sheep (mean starting weight 50.7 6 8.2 kg) received an average doxorubicin dose of 3.8 6 0.5 mg/ kg and were each followed for 6 weeks after their final doxorubicin dose. Systemic adverse effects from doxorubicin including wool loss (n 5 4) and self-limiting diarrhoea (n 5 2), were generally mild. Hematologic monitoring demonstrated stable hemoglobin (94.4 6 8.2 g/L pre-first dose versus 94.5 6 1.7 g/L post-final dose; P 5 NS) and platelet counts (201.0 6 54.0 109/L versus 215.0 6 31.2 109/L post-final dose; P 5 NS), but significant reduction in total white cell count 5.4 6 1.1 109/L versus 3.3 6 0.7 109/L; P 5 .02). Electrocardiographic Monitoring and Cardiac Enzymes
Electrocardiographic monitoring commonly demonstrated transient ST changes during intracoronary drug administration, although infusions were never associated with ectopy or arrhythmia. Heart rate, arterial blood pressure, and left ventricular systolic pressure remained consistent during dosing. Troponin-T values became elevated (O0.1 mg/L) after 53% of all doses (0% positive after first dose, 60% after second dose, 75% after third dose, 100% after fourth dose). Despite this, no hearts had evidence of myocardial infarction, either by CMR or histologic analysis. Late-onset sinus bradycardia was observed in 1 sheep and was persistent during its final CMR and euthanasia studies. Cardiac Function
High-quality electrocardiographic signals were able to be achieved in all animals during CMR imaging, although some positional changes of leads were required to minimize T-wave peaking within the scanner. Doxorubicin resulted in mean left ventricular ejection fractions reduction of 14.9% (range 6.0e28.7%) meeting our aim of inducing moderatesevere cardiac dysfunction (Table 1, Fig. 1). LV systolic dysfunction was global. Overt signs of heart failure were observed in 2 sheep during the later stages of doxorubicin dosing, necessitating frusemide. Right ventricular ejection Table 1. Summary of Results Baseline Weight 50.7 6 8.2 Heart rate beat3/min 103.7 6 10.6 PCWP (mm Hg) 7.8 6 2.3 Cardiac output(L/min) 3.8 6 0.6 Trans thoracic echocardiogram parameters LVEDD (mm) 43.4 6 5.1 LVESD (mm) 29.4 6 3.3 LVFS (%) 31.6 6 1.8 Cardiac magnetic resonance parameters LVEDV (ml) 68.5 6 14.0 LVESV (ml) 37.1 6 8.8 LVEF (%) 46.2 6 4.7 RVEF (%) 39.5 6 5.6
Final
Pvalue
6 6 6 6
8.7 4.5 1.9 0.4
NS NS NS !.05
43.8 6 2.4 35.8 6 2.9 18.2 6 3.9
NS !.005 !.01
6 6 6 6
NS NS !.001 !.05
49.6 105.1 8.0 3.0
69.3 47.7 31.3 28.9
15.8 13.2 8.5 9.6
Baseline and final results are presented as mean 6 standard deviation. LV, left ventricle; EDD, end-diastolic dimension; ESD, end-systolic dimension; FS, fractional shortening; EDV, end-diastolic volume; ESV, endsystolic volume; EF, ejection fraction; RV, right ventricle; NS, not significant.
fraction also significantly decreased by 10.6% from baseline (range 0e26.1%). Additional measures of doxorubicininduced systolic dysfunction included significant reduction in left ventricular fractional shortening and cardiac output. Hemodynamic measurements, including heart rate, arterial blood pressure, pulmonary capillary wedge pressure, and right-sided heart pressures did not significantly change during the study period. Although doxorubicin therapy resulted in increases in left ventricular end-systolic dimension, end-diastolic dilatation was not evident 6 weeks after completion of dosing. CMR assessment of myocardial viability did not show evidence of subendocardial or transmural, delayed enhancement of gadolinium in any sheep, indicating absence of myocardial infarction. However, 2 sheep had gadolinium enhancement in a mid-mural distribution, consistent with nonischemic injury (Fig. 2). Pathology
At necropsy, all treated hearts showed fibrinous adhesions to the pericardium and multifocal areas of pallor, both on epicardial surfaces and irregularly distributed on cut surfaces throughout the myocardium. Microscopically, in right and left ventricles, there was multifocal degeneration and necrosis of cardiac myocytes. Myocyte damage was frequently focal and well-circumscribed (Fig. 4C,D) but sometimes, especially in subendocardial sites, extensive (Fig. 3C). Affected myocytes were either markedly vacuolated (‘‘adria’’ cells) or showed coagulative necrosis characterized by swollen, granular or homogeneous, weakly eosinophilic cytoplasm, loss of striations, and nuclear hyperchromasia (Fig. 3A,B, Fig. 4F). There was substantial edema (confirmed by immunostaining for endogenous albumin) in selected foci of myocardial degeneration (Fig. 4A,B) and frequently a lymphoplasmacytic infiltrate (Fig. 3C, Fig. 4E). Anitchkov cells, with their characteristic ‘‘caterpillar’’ nuclei formed by a wavy ribbon of chromatin (Fig. 3D), were often found in areas of myocardial degeneration. Although the precise origin of these cells is uncertain, they probably represent either macrophages or an abortive attempt at myocyte regeneration. Notably in hearts that had received a fourth dose of doxorubicin, there was substantial replacement fibrosis of cardiomyocytes, well illustrated by Masson’s trichrome staining (Fig. 4C,D). A novel finding was frequent and severe, multivacuolation of Purkinje cells (Fig. 5C), sometimes attended by a lymphocytic infiltrate (Fig. 5D). Vascular changes, in arterioles and venules, were principally characterized by either a nonsuppurative vasculitis with medial myocyte degeneration, pyknosis of endothelial nuclei, and perivascular and mural lymphocytic infiltration (Fig. 5A), or severe, full-thickness degeneration of vessel walls (without an inflammatory reaction) and marked perivascular edema (Fig. 5B). In a few vessels, there was prominent proliferation of vasa vasorum. No thrombi were found in any cardiac blood vessels.
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
789
Fig. 1. (A) Example of steady-state free precession based cardiac magnetic resonance sequences (TrueFISP sequences). Images before and 6 weeks after completion of doxorubicin dosing. Top panel corresponds to pre-doxorubicin (left ventricular ejection fraction [LVEF] 48.5%, left ventricular end-diastolic volume [LVEDV] 53.9 mL, left ventricular end-systolic volume [LVESV] 27.8 mL). Long axis end-diastole (A) and end-systole (B) and short axis end-diastole (C) and end-systole (D). Bottom panel corresponds to post-doxorubicin (LVEF 32.1%, LVEDV 58.8 mL, LVESV 39.9 mL). Long axis end-diastole (E) and end-systole (F) and short axis end-diastole (G) and end-systole (H). LV, left ventricle; RV, right ventricle. (B) All sheep had reduction of LVEF 6 weeks after completion of doxorubicin dosing.
Discussion Animal Models of Cardiomyopathy
In this study, we describe an ovine model of anthracycline-induced cardiac dysfunction using a combination of CMR, hemodynamic, and histopathologic indices. Despite the development of numerous animal models of cardiomyopathy (CMP), none have quite fulfilled all of the criteria necessary to sufficiently mimic human heart failure. Animals used in preclinical research are typically young adults
that are not suffering from the various chronic health problems that often accompany heart failure in elderly human patients. In addition, many experimental protocols used to inflict damage to the heart do so abruptly giving rise to rapid changes in cardiac structure and function, rather than the slowly progressive remodeling that typifies the usual clinical syndrome of CHF. Nevertheless, animal studies are indispensable for complementing in vitro scientific research to provide crucial information relating to the pathophysiology and treatment of heart failure. Although the
790 Journal of Cardiac Failure Vol. 14 No. 9 November 2008
Fig. 2. Delayed myocardial enhancement of gadolinium on cardiac magnetic resonance images in a mid-mural distribution, consistent with nonischemic change. Long axis 4-chamber (A) and short axis mid-ventricular (B). Macroscopic appearance of the same heart specimen at necropsy showing mid-wall myocardial pallor (arrows) (C) and magnified view (D).
importance of small animal research is unquestionable, the successful development of new therapeutic strategies also requires evaluation and refinement in large animals that more closely resemble human anatomy, physiology, and function. Nonischemic, dilated cardiomyopathy (DCM) accounts for a substantial portion of patients with symptomatic cardiac dysfunction, and may arise from numerous etiologies.4,16 Its hallmark is an increase in left ventricular chamber radius-towall thickness ratio with resultant increases in left ventricular wall stress. As with its ischemic counterpart, there is a range of different approaches to inducing preclinical DCM in small and large animal species. Spontaneous models of disease are well recognized in the Syrian hamster17 and in large and giant breed dogs.18 Other well-described models have incorporated experimental tachycardia,19,20 hypertension,21 or mitral regurgitation.22 Chronic pacing tachycardia has emerged as a common method by which to induce DCM and has been applied to sheep, pigs, and dogs.20 Sustained
atrial or ventricular pacing at heart rates between 210 to 240 beats/min characteristically results in biventricular dilatation with reduced cardiac output.19,23 Characteristic pathophysiologic changes include contractile dysfunction of isolated myocytes, abnormal intracellular calcium regulation, abnormalities in b-adrenergic responsiveness, and depletion of high-energy phosphates. Despite advantages, which include a progressive and predictable degree of left ventricular dilation and pump dysfunction and neurohormonal activation, rapid pacing models do not manifest the full spectrum of CHF. Structural and histologic changes can be quite divergent from those observed in clinical CMP and can partly recover on cessation or interruption of pacing.24 In addition, pacing models are not ideal for analysis by CMR, which is now considered a ‘‘gold standard’’ in myocardial assessment for both preclinical and clinical studies. CMR provides superior tissue resolution compared to other noninvasive imaging modalities and can be used to obtain multidimensional information during single studies.12
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
791
Fig. 3. Sheep; ventricle. Large focus of myocardial degeneration (A) with higher power (B) showing severely vacuolated myocytes (black arrow) and myocytes undergoing coagulative necrosis (white arrow). Extensive area of subendocardial degeneration and fibroplasia (C) with lymphoplasmacytic infiltration. Anichkov cells (arrow) with ‘‘caterpillar’’ nuclei (D). (Hematoxylin and eosin).
Doxorubicin-Induced Cardiomyopathy
Toxic cardiomyopathy is an alternative model of nonischemic heart failure that does not have some of the limitations inherent in the pacing-tachycardia strategy. Doxorubicin (Adriamycin) has traditionally been the ‘‘toxin of choice’’ across different animal species. This quinonecontaining anthracycline antibiotic is most frequently used as a chemotherapeutic agent for various types of solid tumors and hematologic malignancies. Its major clinical shortcoming is dose-related cardiotoxicity, which can manifest both as acute myocarditis and progressive, chronic CMP. In human patients, cardiotoxicity is largely determined by cumulative dose exposure, such that life-threatening CHF effects 10% of patients who have received more than 550 mg/m2 systemic dose.25 The mechanistic pathways mediating such toxicity are multiple and complex, with damage to mitochondrial respiratory chain function and the liberation of highly reactive, free oxygen radicals playing a key role in initiating a cascade of deleterious subcellular sequelae. Cardiomyocyte morphological changes include cytoplasmic vacuolation and myofibrillar loss, with the end-result being apoptosis of both myocytes and other cardiac-relevant cells, such as endothelial cells and replacement fibrosis.9,26 Although several groups have reported successful induction of cardiac dysfunction in animals administered this drug, their strategies have differed considerably especially
with respect to dosing regimen and route of drug administration. In small animal models of myocarditis and heart failure, doxorubicin has usually been administered by either single intraperitoneal dose5 or serial intravenous doses.7,27 Repeat intravenous administration has also been used in dogs,28 goats, and sheep, sometimes in combination with arteriovenous anastomosis29,30; however, the high doses of medication required have been associated with substantial systemic effects and mortality. The coronary route of administration therefore has been applied to enable lower doses to be given, with high firstpass concentrations in the myocardium and fewer systemic effects.11,31e33 In the early canine models, dosing was typically weekly for up to 5 doses and oral verapamil therapy was sometimes combined with doxorubicin exposure to augment cardiotoxicity.32,34 Mortality was still substantial (20e60%), often because of fatal arrhythmia. The first description of doxorubicin-induced cardiomyopathy by coronary administration in sheep was reported recently by Borenstein et al,11 who partly attributed their improvement in mortality rate to the creation of a pericardial window, which was designed to prevent doxorubicin-induced inflammatory pericardial effusion and tamponade. Assessment of cardiac dysfunction in this study was both by echocardiography, including tissue Doppler imaging, and left ventriculography. Animals received a mean of 2.5 doses of 0.75
792 Journal of Cardiac Failure Vol. 14 No. 9 November 2008
Fig. 4. Sheep; ventricle. Focal myocardial degeneration and cardiomyocyte loss (A) with edematous area immunostained for endogenous albumin (B) showing marked albumin extravasation. Well-circumscribed focus of myocardial damage (C) with substantial reparative fibrosis evident by Masson’s trichrome staining (D). Focal myocardial injury showing lymphoplasmacytic infiltrate and residual degenerate myocytes (E) and same area stained with Masson’s trichrome showing effete myocytes marooned in collagenous connective tissue (F).
mg/kg, resulting in a mean ejection fraction decline of approximately 16%. In our pilot work, using Merino sheep, initial dosing with 0.75 mg/kg dose did not achieve the intended deterioration in left ventricular contractility. Consequently, doses were increased to 1 mg/kg and sheep required more doses than was the case in the Borenstein description of this model. The nonsurvival rate we observed (20%) was favorable compared with the range reported by earlier studies, with both deaths in our group likely to be due to arrhythmia, accompanied by acute pulmonary edema. Surviving sheep demonstrated relatively mild systemic side effects and weight was well maintained throughout the full study protocol.
Unlike swine, the coronary anatomy of sheep is not as similar to humans, with sheep demonstrating a left dominant circulation with large circumflex artery. In our experience, the left main coronary artery in Merino sheep is typically short and difficult to maintain stable catheter engagement over a prolonged period. Consequently, we selectively engaged and dosed the left anterior descending and left circumflex arteries in sequence, using an AL I catheter. Although high-dose doxorubicin (1 mg/kg) was administered over a shorter period than in other studies, we did not observe intraoperative arrhythmia and did not lose any sheep directly from intraprocedural complications. Heart rate, blood pressure, and left ventricular systolic
Ovine Model of Toxic, Nonischemic Cardiomyopathy
Psaltis et al
793
Fig. 5. Sheep; ventricle. Nonsuppurative vasculitis (A) with lymphocytic infiltration of degenerate wall (arrow) and severe vasculopathy (B) showing dissolution of the vessel wall and marked perivascular edema. Severe vacuolation (arrow) of Purkinje cells (C) with higher power view as inset and a lymphocytic inflammatory reaction (D). (Hematoxylin and eosin).
pressure did not fluctuate substantially during individual doses and, although ST depression was occasionally observed, this phenomenon was transient and almost never residual at the end of surgery. Evaluation by CMR
This is the first description of the utility of CMR in a preclinical model of doxorubicin-induced CMP. Although we used serial TTE to monitor the progressive nature of cardiac decline and to indicate which sheep might require a fourth dose of doxorubicin, echocardiographic image quality was variable, such that right parasternal images were not always satisfactorily obtainable. Magnetic resonance imaging provided superior resolution than TTE for the evaluation of cardiac geometry and function. It also provided a noninvasive measure of viability, to complement histopathology in ruling out transmural myocardial infarction, through the assessment of delayed gadolinium enhancement. Two of 8 sheep displayed patchy, mid-mural uptake of gadolinium, which closely mirrors the pattern and prevalence of late gadolinium uptake reported in human patients with DCM.35,36 The exclusion of myocardial infarction was especially important given the frequency of elevated troponin-T levels after doxorubicin doses and led us to infer that this was probably the result of progressive doxorubicin-induced myocarditis, rather than infarction.
Although there was some individual variation in response to high-dose intracoronary doxorubicin, all sheep had global deterioration in left ventricular function. In addition, we also observed reduction in mean right ventricular ejection fraction, not reported by other studies in which doxorubicin has been administered without adjuvant arteriovenous fistula creation.30 Notably, we did not achieve end-diastolic dilation of the left ventricle, contrasting the previous report by Borenstein et al, who adopted a lower dose protocol, with fewer doses and shorter follow-up, but obtained their measurements by 2-dimensional echocardiography.11 In future studies, we intend to follow animals for longer periods to determine whether dilation becomes a later feature of our model. The pathologic findings observed were consistent with those of previous studies and with the changes reported in human patients suffering anthracycline-induced cardiomyopathy. Lesions were characterized by either multifocal degeneration and necrosis of cardiomyocytes with attendant inflammation or myocyte loss and replacement fibrosis, and nonocclusive vasculopathy. It has been previously demonstrated that these doxorubicin-induced effects are largely irreversible and progressive over time, resulting in a more suitable pathologic substrate for preclinical trials of heart failure therapy (eg, antifibrotic drugs and cellular cardiomyoplasty) than traditional pacing models. Another notable
794 Journal of Cardiac Failure Vol. 14 No. 9 November 2008 finding of our study was the consistent damage caused by doxorubicin to Purkinje fibres. This has not been described in previous in vivo reports, although anthracyclines are known to affect Purkinje fibre action potentials in vitro.37 Such changes may contribute to the substantial rate of arrhythmia known to complicate anthracycline cardiomyopathy and will be the subject of future electrophysiologic testing in our model.
6.
7.
8.
Study Limitations
A limitation of our description of doxorubicin cardiotoxicity is the lack of invasive left ventricular hemodynamic analysis by conductance catheters and the absence of assessment of myocardial energo-mechanics. The sequential intracoronary dosing regime requires multiple repeat anesthetics and measurements performed in the anesthetized state may differ from those at conscious and closechest state. Although we were satisfied by the progressive response of cardiac systolic function to doxorubicin, there inevitably was some variability in the degree of response between individual sheep and this supports the need to titrate the number of doses to obtain the desired effect.
9.
10.
11.
12.
13.
Conclusion In conclusion, we describe an ovine model of moderatesevere biventricular dysfunction, induced by repeat coronary administration of doxorubicin. Moreover, we demonstrate the preclinical utility of CMR for noninvasive and comprehensive assessment of cardiac anatomy, function, and viability. This information, coupled by histopathologic analysis, has shown this model to be representative of nonischemic cardiomyopathy seen in humans.
14.
15.
16. 17.
Acknowledgment The authors would also like to thank Dr Tim Kuchel, Ms Jodie Dier, Ms Melissa Gourlay, and Mr Adrian Hines (Veterinary Services Division, IMVS); Ms Kerry Williams and Mr Leigh Penney (Perrett Medical Imaging, Wakefield Street, Adelaide); and Dr Zhao Cai (Hanson Institute Centre for Neurological Diseases, IMVS) for their assistance during this study.
18.
19.
20. 21.
References 22. 1. Fonarow GC. Heart failure: recent advances in prevention and treatment. Rev Cardiovasc Med 2000;1:25e33, 54. 2. Jessup M., Brozena S. Heart failure. N Engl J.M. 2003;348: 2007e2018. 3. Laflamme MA, Murry CE. Regenerating the heart. Nat Biotechnol 2005;23:845e56. 4. Follath F. Ischemic versus non-ischemic heart failure: should the etiology be determined? Heart Fail Mon 2001;1:122e5. 5. Agbulut O, Menot ML, Li Z, Marotte F, Paulin D, Hagege AA, et al. Temporal patterns of bone marrow cell differentiation following
23.
24.
transplantation in doxorubicin-induced cardiomyopathy. Cardiovasc Res 2003;58:451e9. Borenstein N, Chetboul V, Bruneval P, Hekmati M, Tissier R, Behr L, et al. Non-cultured cell transplantation in an ovine model of nonischemic heart failure. Eur J Cardiothorac Surg 2007;31:444e51. Aupperle H, Garbade J, Schubert A, Barten M, Dhein S, Schoon HA, et al. Effects of autologous stem cells on immunohistochemical patterns and gene expression of metalloproteinases and their tissue inhibitors in doxorubicin cardiomyopathy in a rabbit model. Vet Pathol 2007;44:494e503. Yarbrough WM, Spinale FG. Large animal models of congestive heart failure: a critical step in translating basic observations into clinical applications. J Nucl Cardiol 2003;10:77e86. Takemura G, Fujiwara H. Doxorubicin-induced cardiomyopathy from the cardiotoxic mechanisms to management. Progress Cardiovasc Dis 2007;49:330e52. Swain SM, Whaley FS, Ewer MS. Congestive heart failure in patients treated with doxorubicin: a retrospective analysis of three trials. Cancer 2003;97:2869e79. Borenstein N, Bruneval P, Behr L, Laborde F, Montarras D, Daures JP, et al. An ovine model of chronic heart failure: echocardiographic and tissue Doppler imaging characterization. J Cardiac Surg 2006;21: 50e6. Teo KS, Carbone A, Piantadosi C, Chew DP, Hammett CJ, Brown MA, et al. Cardiac MRI assessment of left and right ventricular parameters in healthy Australian normal volunteers. Heart Lung Circ 2008;17:313e7. Kim RJ, Wu E, Rafael A, Chen EL, Parker MA, Simonetti O, et al. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med 2000;343: 1445e53. McCrohon JA, Moon JC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJ, et al. Differentiation of heart failure related to dilated cardiomyopathy and coronary artery disease using gadoliniumenhanced cardiovascular magnetic resonance. Circulation 2003; 108:54e9. Jerosch-Herold M, Seethamraju RT, Swingen CM, Wilke NM, Stillman AE. Analysis of myocardial perfusion MRI. J Magn Reson Imaging 2004;19:758e70. Eriksson H. Heart failure: a growing public health problem. J Intern Med 1995;237:135e41. Gertz EW, Stam AC Jr, Sonnenblick EH. A quantitative and qualitative defect in the sarcoplasmic reticulum in the hereditary cardiomyopathy of the Syrian hamster. Biochem Biophys Res Communi 1970;40:746e53. Tidholm A, Svensson H, Sylven C. Survival and prognostic factors in 189 dogs with dilated cardiomyopathy. J Am Animal Hosp Assoc 1997;33:364e8. Yamamoto K, Burnett JC Jr, Meyer LM, Sinclair L, Stevens TL, Redfield MM. Ventricular remodeling during development and recovery from modified tachycardia-induced cardiomyopathy model. Am J Physiol 1996;271:R1529e34. Recchia FA, Lionetti V. Animal models of dilated cardiomyopathy for translational research. Vet Res Commun 2007;31(Suppl 1):35e41. Aoyagi T, Mirsky I, Flanagan MF, Currier JJ, Colan SD, Fujii AM. Myocardial function in immature and mature sheep with pressureoverload hypertrophy. Am J Physiol 1992;262:H1036e48. Nakano K, Swindle MM, Spinale F, Ishihara K, Kanazawa S, Smith A, et al. Depressed contractile function due to canine mitral regurgitation improves after correction of the volume overload. J Clin Invest 1991; 87:2077e86. Tanaka R, Spinale FG, Crawford FA, Zile MR. Effect of chronic supraventricular tachycardia on left ventricular function and structure in newborn pigs. J Am Coll Cardiol 1992;20:1650e60. Spinale FG, Holzgrefe HH, Mukherjee R, Arthur SR, Child MJ, Powell JR, et al. LV and myocyte structure and function after early recovery from tachycardia-induced cardiomyopathy. Am J Physiol 1995; 268:H836e47.
Ovine Model of Toxic, Nonischemic Cardiomyopathy 25. Singal PK, Iliskovic N. Doxorubicin-induced cardiomyopathy. N Engl J Med 1998;339:900e5. 26. Kalyanaraman B, Joseph J, Kalivendi S, Wang S, Konorev E, Kotamraju S. Doxorubicin-induced apoptosis: implications in cardiotoxicity. Mol Cell Biochem 2002;234-235:119e24. 27. Schwarz ER, Pollick C, Dow J, Patterson M, Birnbaum Y, Kloner RA. A small animal model of non-ischemic cardiomyopathy and its evaluation by transthoracic echocardiography. Cardiovasc Res 1998;39: 216e23. 28. Ditchey RV, LeWinter MM, Higgins CB. Acute effects of doxorubicin (Adriamycin) on left ventricular function in dogs. Int J Cardiol 1984;6: 341e53. 29. Tessier D, Lajos P, Braunberger E, Pouchelon JL, Carpentier A, Chachques JC, et al. Induction of chronic cardiac insufficiency by arteriovenous fistula and doxorubicin administration. J Cardiac Surg 2003;18:307e11. 30. Chekanov VS. A stable model of chronic bilateral ventricular insufficiency (dilated cardiomyopathy) induced by arteriovenous anastomosis and doxorubicin administration in sheep. J Thorac Cardiovasc Surg 1999;117:198e9.
Psaltis et al
795
31. Toyoda Y, Okada M, Kashem MA. A canine model of dilated cardiomyopathy induced by repetitive intracoronary doxorubicin administration. J Thorac Cardiovasc Surg 1998;115:1367e73. 32. Monnet E, Orton EC. A canine model of heart failure by intracoronary Adriamycin injection: hemodynamic and energetic results. J Cardiac Fail 1999;5:255e64. 33. Magovern JA, Christlieb IY, Badylak SF, Lantz GC, Kao RL. A model of left ventricular dysfunction caused by intracoronary Adriamycin. Ann Thorac Surg 1992;53:861e3. 34. Bright JM, Buss DD. Effects of verapamil on chronic doxorubicininduced cardiotoxicity in dogs. J Natl Cancer Inst 1990;82:963e4. 35. Stork A, Mullerleile K, Bansmann PM, Koops A, Meinertz T, Adam G, et al. [Patterns of delayed-enhancement in MRI of ischemic and non-ischemic cardiomyopathies]. Rofo 2007;179:21e30. 36. Gottlieb I, Macedo R, Bluemke DA, Lima JA. Magnetic resonance imaging in the evaluation of non-ischemic cardiomyopathies: current applications and future perspectives. Heart Fail Rev 2006;11:313e23. 37. Rabkin SW, Bose D. Adriamycin induced alterations in canine Purkinje fiber action potential. Res Commun Chem Pathol Pharmacol 1981;34:55e67.