Objective Anthracyclines such as doxorubicin are an effective class of antineoplastic agents. Despite its efficacy in the treatment of a variety of cancers, the clinical use of doxorubicin is limited by cardiac side effects. While it has been suggested that doxorubicin alters myocardial fatty acid metabolism, it is poorly understood whether this is the case and whether variations in myocardial triacylglycerol (TAG) metabolism contribute to doxorubicin induced cardiotoxicity. Since TAG catabolism in the heart is controlled by adipose triglyceride lipase (ATGL), this study examined the influence of doxorubicin on cardiac energy metabolism and TAG values as well as the consequence of forced expression of ATGL in the setting of doxorubicin induced cardiotoxicity.
Design and setting Wild type (WT) mice and mice with cardiomyocyte specific ATGL overexpression were divided into two groups per genotype that received a weekly intraperitoneal injection of saline or doxorubicin for 4 weeks.
Results Four weeks of doxorubicin administration significantly impaired in vivo systolic function (11% reduction in ejection fraction, p<0.05), which was associated with increased lung wet to dry weight ratios. Furthermore, doxorubicin induced cardiac dysfunction was independent of changes in glucose and fatty acid oxidation in WT hearts. However, doxorubicin administration significantly reduced myocardial TAG content in WT mice (p<0.05). Importantly, cardiomyocyte specific ATGL overexpression and the resulting decrease in cardiac TAG accumulation attenuated the decrease in ejection fraction (p<0.05) and thus protected mice from doxorubicin induced cardiac dysfunction.
Conclusions Taken together, our data suggest that chronic reduction in myocardial TAG content by cardiomyocyte specific ATGL overexpression is able to prevent doxorubicin induced cardiac dysfunction.
Statistics from Altmetric.com
Anthracyclines such as doxorubicin are an effective class of antineoplastic agents for the treatment of many human neoplasms, including breast cancer.1 However, the clinical use of doxorubicin is limited by cardiac side effects.2 Doxorubicin administration is known to cause cardiotoxicity that can develop into congestive heart failure.2 At present, medical management of heart failure due to doxorubicin administration is similar to that of all patients with left ventricular (LV) dysfunction and heart failure2 and does not specifically target doxorubicin induced cardiotoxicity. This is largely due to our lack of understanding of the mechanisms that lead to doxorubicin induced cardiac dysfunction. Therefore, identification of novel pharmacologic targets and development of medical treatments that aid in the management of doxorubicin induced cardiac dysfunction is paramount to the improvement of patient care and the long term survival of these individuals.
Although several mechanisms have been proposed for doxorubicin induced cardiac dysfunction, including enhanced oxidative stress, increased topoisomerase II activity, apoptosis, and collagen synthesis/degradation,3 targeting these pathways has yielded insufficient success in improving doxorubicin induced cardiomyopathy.4 Recent studies have associated alterations in myocardial energy metabolism with the cardiotoxic effects of doxorubicin.5 These studies have demonstrated that overall oxidative phosphorylation in the cardiomyocyte and adenosine triphosphate (ATP) production is decreased with doxorubicin administration. Furthermore, previous studies have suggested that myocardial substrate utilisation is also altered with doxorubicin administration and that fatty acid oxidation decreases with doxorubicin induced cardiotoxicity.6 However, to date, no studies have examined the dynamic myocardial triacylglycerol (TAG) pool as it pertains to substrate utilisation during doxorubicin induced cardiotoxicity. This is increasingly important since there is growing evidence that promotion of a hypolipidaemic cardiomyocyte environment diminishes the cardiotoxic effects of doxorubicin treated animals and patients.7 As well, there is increasing evidence for the importance of the intramyocardial TAG metabolism as it pertains to overall cardiac energy metabolism in physiology and pathophysiology.8
Despite these reported alterations in energy metabolism in doxorubicin induced cardiomyopathy, the reported benefits of a hypolipidaemic cardiomyocyte environment in diminishing the cardiac side effects of doxorubicin, and the critical role of TAG turnover in regulating overall fatty acid utilisation, no studies have specifically addressed whether myocardial TAG metabolism and accumulation are altered following doxorubicin administration, and whether manipulating cardiac TAG metabolism could modify the natural history of doxorubicin induced cardiac dysfunction. To examine this, the aims of this study were to: (1) investigate the effects of doxorubicin administration on myocardial exogenous substrate metabolism and TAG metabolism; and (2) determine if manipulation of the intramyocardial TAG pool could be used as an approach to treat doxorubicin induced cardiac dysfunction. As TAG catabolism in the heart is primarily controlled by adipose triglyceride lipase (ATGL),9 this study examined whether maintaining a chronically low intramyocardial TAG pool, via forced expression of ATGL specifically in cardiomyocytes, influences doxorubicin induced cardiotoxicity. Our results show that maintaining a state of chronically low intramyocardial TAGs with concurrent administration of doxorubicin is sufficient to protect from doxorubicin induced cardiac dysfunction.
ATGL overexpressing (MHC-ATGL) mice were generated as described previously.9 A detailed description is presented in the online supplements. Female mice were chosen since female breast cancer is the most common cancer in North America and male breast cancer is a clinically rare entity.10 Thus the study of cardiotoxic effects of doxorubicin is most appropriately conducted in female animals.
Induction of doxorubicin induced cardiac dysfunction
Mice were divided into four groups: (1) wild type (WT)+saline (n=5); (2) WT+doxorubicin (n=7); (3) MHC-ATGL+saline (n=5); and (4) MHC-ATGL+doxorubicin (n=7). Doxorubicin (Sigma) was administered weekly at a cumulative dose of 32 mg/kg via intraperitoneal injections at 8 mg/kg bodyweight for 4 weeks. The corresponding volume of saline was administered to the control groups. Previous reports have shown LV dysfunction without overt heart failure with this doxorubicin dose.11 Thus we chose this protocol to mimic a clinically relevant scenario since the incidence of heart failure in recent doxorubicin regimen trials is <2.1%.12
Transthoracic echocardiography was performed as described previously.9 A detailed description is presented in the online supplements.
Hearts were aerobically perfused in the working mode as previously described.9 A detailed description is presented in the online supplements.
Tissue homogenisation and lipid analysis
Tissue homogenisation, lysate protein assay, and tissue TAG content analysis was performed as described previously.9 Quantification of long chain acyl coenzyme A (acyl-CoA) species and ceramides was performed by ultra-performance liquid chromatography (UPLC).9 A detailed description is presented in the online supplements.
Immunoblot analysis was performed as described previously.9 A detailed description is presented in the online supplements.
Serum triacylglycerol analysis
Serum TAG concentration was determined using the 2780-400H Infinity TAG reagent (Thermo MA).
Gene expression analysis
Gene expression analysis was performed using quantitative reverse transcriptase PCR.9 A detailed description is presented in the online supplements.
Results are expressed as means±SEM. Statistical analyses were performed using GraphPad Prism software. Comparisons between two groups were made by unpaired two tailed Student's t test. Values of p<0.05 were considered statistically significant.
Four weeks of doxorubicin administration impairs in vivo systolic function but not diastolic function in WT mice
Doxorubicin administration of 8 mg/kg intraperitoneally once a week for four consecutive weeks resulted in no observed mortality. While bodyweights were unchanged following saline administration in WT mice, a significant 8.9% loss in bodyweight was observed following doxorubicin administration (figure 1A), suggesting that this doxorubicin treatment regimen elicited a systemic response similar to previous reports.13 We next determined whether 4 weeks of doxorubicin treatment led to morphological changes in the heart of WT mice. Indeed, doxorubicin treated WT mice exhibited significant cardiac atrophy when compared to the saline treated counterparts, as was indicated by the decreased ratio of ventricular weight to tibia length in doxorubicin treated WT mice (figure 1B). To assess further the effect of doxorubicin treatment on ventricular wall dimensions, we performed non-invasive transthoracic echocardiography. Consistent with the observed morphological changes, LV posterior wall thickness in systole (figure 1C) and diastole (figure 1D) was significantly decreased in WT mice following doxorubicin administration, suggesting that 4 weeks of doxorubicin treatment causes significant thinning of the ventricular free walls. Importantly, echocardiographic analysis also showed an 11% reduction in the ejection fraction of doxorubicin treated WT mice, suggesting systolic dysfunction (figure 1E,F). This decrease in systolic function was associated with a significant increase in the ratio of wet to dry lung weight (figure 1G) in WT mice treated with doxorubicin compared to the saline treated controls. In contrast to the changes in systolic function, parameters of diastolic function such as mitral valve deceleration time, E/A ratio, E/E’ and s/d ratio were unchanged between WT mice treated with doxorubicin and WT mice treated with saline (table 1). Taken together, these findings suggest that following 4 weeks of doxorubicin treatment, there is cardiac atrophy accompanied by a parallel selective decline in systolic function.
Doxorubicin induced cardiac dysfunction in vivo is independent of changes in glucose and fatty acid oxidation in WT hearts
Previous work using neonatal rat cardiomyocytes, animal models, and human studies have suggested that doxorubicin administration decreases cardiomyocyte overall myocardial energy production by reducing oxidative phosphorylation rates.14 However, this has not been directly measured in the working heart. Therefore, to determine whether alterations in myocardial energy metabolism could contribute to doxorubicin induced cardiac dysfunction in WT mice, we measured myocardial glucose and oleate oxidation. This was accomplished by subjecting hearts to ex vivo perfusions in the working mode using radiolabelled substrates. Since heart rate (figure 2A) and LV minute work (figure 2B) ex vivo were comparable between hearts from doxorubicin and saline treated WT mice, differences in substrate metabolism could be evaluated in the absence of alterations in energetic demand. In contrast to previous reports,14 our data show that rates of glucose (figure 2C) and oleate oxidation (figure 2D) were unchanged in ex vivo perfused working hearts from WT mice following doxorubicin treatment. Taken together, these findings demonstrate that 4 weeks of doxorubicin treatment does not lead to changes in mitochondrial oxidative metabolism of glucose and fatty acids, and that doxorubicin induced cardiac dysfunction in WT hearts is independent of changes in oxidative substrate metabolism.
Doxorubicin administration reduces myocardial TAG content in WT mice
Since it has been recently shown that doxorubicin treatment causes cardiac lipotoxicity,15 we next examined whether changes in myocardial accumulation of toxic fatty acid metabolites and TAG could contribute to doxorubicin induced cardiotoxicity. Interestingly, 4 weeks of doxorubicin administration did not lead to alterations in total long chain acyl-CoA (figure 3A) and ceramide (figure 3B) content in hearts from WT mice when compared to the saline treated controls. In addition, the myocardial mRNA expression of enzymes involved in the synthesis of TAG in the myocardium [1-acyl-sn-glycerol-3-phosphate acyltransferase (aGPAT) and diacylglycerol acyltransferase-2 (DGAT2)], were not decreased with doxorubicin treatment in any of the groups (data not shown). However, doxorubicin administration significantly decreases serum TAG levels in WT mice (figure 3C). This finding is consistent with the significant decrease in intramyocardial TAG content in hearts from doxorubicin treated WT mice (figure 3D). Interestingly, activating phosphorylation of hormone sensitive lipase (HSL) at Serine 660, an enzyme involved in TAG catabolism, was similar between WT mice treated with saline and doxorubicin (figure 3E). In addition, protein expression of the lipid droplet coat protein, perilipin 5 (PLIN5), which has been shown to regulate TAG storage and catabolism,16 remained unchanged in hearts from WT mice following doxorubicin administration (figure 3F). Moreover, ATGL mRNA expression was unchanged (figure 3G) and there was an non-statistically significant trend towards upregulation of ATGL protein expression in hearts from WT mice treated with doxorubicin (figure 3H). Together, these data suggest that cardiac dysfunction following 4 weeks of doxorubicin administration is associated with a reduction in myocardial TAG concentrations and is independent of alterations in known lipotoxic species such as long chain acyl-CoAs and ceramides.
Cardiomyocyte specific ATGL overexpression protects from doxorubicin induced cardiac dysfunction
Since doxorubicin induced cardiomyopathy was associated with a significant decrease in circulating TAG levels and myocardial TAG content (figure 3C), we hypothesised that altering the intramyocardial TAG pool could influence the functional outcome of doxorubicin induced cardiotoxicity. To test this hypothesis, we utilised mice with cardiomyocyte specific overexpression of ATGL (MHC-ATGL; figure 3E–G). This is a mouse model with chronically reduced myocardial TAG content (figure 3C) and PLIN5 protein expression (figure 3G), as well as unchanged concentrations of long chain acyl-CoAs (figure 3A) and ceramides (figure 3B) at baseline.
Following 4 weeks of doxorubicin administration, MHC-ATGL mice showed a significant 9.5% loss in bodyweight (figure 1A) that was similar to that observed in WT mice, suggesting that the systemic response to doxorubicin treatment was comparable between genotypes. However, when specifically assessing cardiac morphology, we observed that cardiomyocyte specific ATGL overexpression ameliorated the significant doxorubicin induced cardiac atrophy exhibited in WT mice, as was indicated by the unchanged ratio of ventricular weight to tibia length (figure 1B) and LV posterior wall thickness in systole (figure 1C) and diastole (figure 1D). Importantly, echocardiographic analysis also showed unchanged ejection fraction in doxorubicin treated MHC-ATGL mice, suggesting that ATGL overexpression provided protection from systolic dysfunction following 4 weeks of doxorubicin administration. In addition, this protection from in vivo systolic dysfunction preserved the ratio of wet to dry lung weight (figure 1G) in MHC-ATGL mice treated with doxorubicin, when compared to the saline treated mice. Also, echocardiographic parameters of diastolic function were unchanged in doxorubicin treated MHC-ATGL mice.
Similar to WT hearts, oxidation rates of glucose and oleate were unchanged in ex vivo perfused working hearts form MHC-ATGL mice following doxorubicin treatment (figure 2C,D), corroborating that doxorubicin induced cardiac dysfunction in WT mice and the protection from doxorubicin induced cardiomyopathy in MHC-ATGL mice are not secondary to changes in mitochondrial oxidation of glucose and fatty acids. Taken together, these data suggest that following 4 weeks of doxorubicin treatment, cardiomyocyte specific ATGL overexpression and the resulting decrease in myocardial TAG content are sufficient to protect against doxorubicin induced cardiac atrophy, systolic dysfunction, and pulmonary signs of congestive heart failure.
We investigated whether doxorubicin induced cardiac dysfunction is accompanied by changes in myocardial exogenous substrate metabolism, lipotoxicity, and TAG accumulation. Additionally, we also studied whether manipulation of the intramyocardial TAG pool could be used as a novel method to ameliorate doxorubicin induced cardiac dysfunction. As expected, doxorubicin administration resulted in cardiac atrophy that was paralleled by a selective decline in systolic function and precipitating pulmonary signs of congestive heart failure in WT mice. In contrast, mice that maintained a chronically low intramyocardial TAG pool via forced expression of ATGL were resistant to doxorubicin induced cardiac atrophy and LV systolic dysfunction, and were protected from the pulmonary signs of congestive heart failure. These data show that forced expression of ATGL and a chronically decreased myocardial TAG pool protects against doxorubicin induced cardiomyopathy.
In order to investigate the mechanisms by which increased ATGL expression, leading to a decreased myocardial TAG pool, protects against doxorubicin induced cardiomyopathy, we examined myocardial substrate metabolism. Indeed, an increasing number of studies suggest that alterations in myocardial exogenous substrate metabolism are associated with doxorubicin induced cardiac dysfunction.14 Specifically, previous work has suggested that doxorubicin treatment decreases fatty acid and glucose oxidation and the expression of enzymes involved in fatty acid and glucose metabolism.14 However, it is still unclear if doxorubicin treatment changes myocardial oxidative phosphorylation in the intact working heart and whether this is preceded or followed by changes in cardiac function. Therefore, in order to understand the role that myocardial exogenous substrate metabolism plays in doxorubicin induced cardiac dysfunction, we examined myocardial fatty acid and glucose oxidation rates using ex vivo working heart perfusions. In contrast to previous reports,14 fatty acid and glucose oxidation rates were unchanged and thus not associated with the cardiac dysfunction induced by this doxorubicin treatment protocol in WT and MHC-ATGL mice. This suggests that the doxorubicin induced decrease in cardiac contractility develops independently from changes in mitochondrial oxidation of glucose and fatty acids. These findings also show that the ability of cardiomyocyte specific ATGL overexpression to prevent doxorubicin induced cardiac dysfunction is not due to alterations in cardiac exogenous substrate metabolism or impaired oxidative phosphorylation. Together, these findings are particularly important as they highlight that mechanisms other than impaired myocardial substrate metabolism14 ,17 are contributing to doxorubicin induced cardiac dysfunction. That said, overall doxorubicin damage to the hearts is less severe in the present study compared to previous studies.6 Indeed, the doxorubicin administration protocol used in the present study does not cause overt heart failure but rather cardiac dysfunction, so as to mimic more closely the clinical scenario. Another difference in our model compared to some of the previous work is that we have measured cardiac energy metabolism in the perfused working heart, rather than in isolated cardiomyocytes.6 Therefore, our data suggest that alterations in energy metabolism may occur as a consequence of impaired function as opposed to causing cardiac dysfunction.
As alterations in myocardial exogenous substrate metabolism did not appear to be responsible for doxorubicin induced cardiac dysfunction or the beneficial effects of ATGL overexpression, we examined other possible mechanisms. Recent studies have suggested that doxorubicin treatment leads to systemic lipotoxicity15 and that this may contribute to doxorubicin induced cardiotoxicity. To address this, we examined the effects of doxorubicin administration on intramyocardial long chain acyl-CoA and ceramides content in WT and MHC-ATGL mice, as these are lipotoxic to the cardiomyocyte.18 In contrast to previous work that reported increased accumulation of lipotoxic intermediates with the administration of doxorubicin in cultured cells,19 we found that these lipotoxic fatty acid metabolites were unaltered by doxorubicin administration in hearts from WT mice. Similar to WT mice, MHC-ATGL mice treated with doxorubicin did not show altered accumulation of intramyocardial ceramide content. However, hearts from MHC-ATGL mice treated with doxorubicin showed a significant decrease in long chain acyl-CoA content compared to hearts from saline treated MHC-ATGL mice or doxorubicin treated WT mice. These latter findings suggest that a decrease in myocardial long chain acyl-CoA content induced by overexpression of ATGL may be partially responsible for preventing doxorubicin induced cardiotoxicity and cardiac dysfunction. In fact, our data are entirely consistent with previous work in a rat model of doxorubicin induced cardiotoxicity showing that a reduction of cardiomyocyte lipid accumulation diminishes the cardiac side effects of doxorubicin.7
Although forced expression of ATGL may protect against doxorubicin induced cardiac dysfunction via reduced long chain acyl-CoA content, we also investigated whether alterations in TAG concentrations and/or TAG catabolism may be involved. This is particularly important since there is growing evidence indicating that a hypolipidaemic cardiomyocyte environment diminishes the cardiotoxic effects of doxorubicin treated animals and patients.7 In addition, previous studies have suggested that cardiomyocyte TAG catabolism plays an important role in regulating cardiac function in pathological conditions,8 ,20 further supporting the notion that TAG concentrations may be important in doxorubicin induced cardiotoxicity. Based on this rationale, we assessed intramyocardial TAG content as well as expression or phosphorylation of enzymes involved in TAG catabolism (HSL and ATGL) and storage (PLIN5). Interestingly, doxorubicin treatment led to a significant decrease in intramyocardial TAG content without significant changes to HSL activating phosphorylation at Ser660 or expression of ATGL and PLIN5 in WT mice.
Because of these findings, and the observation that myocardial fatty acid oxidation was unchanged following doxorubicin treatment, it is tempting to speculate that rather than enhanced TAG catabolism, reduced fatty acid uptake and incorporation into newly formed TAG contributes to the decrease in TAG accumulation following doxorubicin treatment. Although this decrease in intramyocardial TAG concentrations in WT mice treated with doxorubicin may suggest that this contributes to cardiac dysfunction, a more dramatic reduction in cardiac TAG values in MHC-ATGL mice was associated with preserved cardiac function. Despite using a sophisticated mouse model that specifically decreases cardiac TAG concentrations, it is possible that other mechanisms are contributing to the protective effects from doxorubicin induced cardiac dysfunction. As such, further studies are required to elucidate these potential mechanisms. However, based on these findings, we propose that the reduction in myocardial TAG accumulation in WT hearts is an insufficient adaptive response to doxorubicin treatment and that more pronounced lowering of intramyocardial TAGs induced by ATGL overexpression contributes to preventing doxorubicin induced cardiac dysfunction. Therefore, the observed decrease of intramyocardial TAG content in hearts treated with doxorubicin is an adaptive response that is further enhanced by ATGL overexpression.
This study assessed the effects of doxorubicin administration on myocardial oxidative metabolism of glucose and fatty acids in the ex vivo perfused working heart as well as TAG accumulation. In contrast to previous reports, our findings demonstrate that doxorubicin induced cardiac dysfunction occurs in the absence of alterations in myocardial glucose and fatty acid oxidation. More importantly, our results show that manipulation of intramyocardial TAG metabolism leading to a chronically decreased myocardial TAG pool protects against doxorubicin induced cardiomyopathy. Moreover, these findings suggest that the reduction in myocardial TAG accumulation in WT hearts is an adaptive, albeit insufficient, response to doxorubicin treatment. Therefore, novel pharmacotherapy aimed at reducing myocardial TAG concentrations may provide new avenues for the treatment of cardiomyopathy in patients with cancer undergoing anthracycline chemotherapy by producing a hypolipid environment within cardiomyocytes.
Acknowledgements The authors acknowledge the expert technical assistance of Amy Barr, Jamie Boisvenue, Donna Becker, Jody Levasseur, Grant Masson, Carrie-Lynn Soltys, and Ken Strynadka.
Contributors JN, PCK, TP and JRBD participated in original conception and design of the study. All authors participated in the interpretation of the data, drafting and critical revisions of the paper and the final manuscript submitted.
Funding This work was supported by grants from the Canadian Institute of Health Research to JRBD, doctoral studentships from the Mazankowski Alberta Heart Institute, Alberta Innovates Health Solutions, and the Canadian Diabetes Association to JN, a postdoctoral fellowship from the Heart and Stroke Foundation of Canada and the Canadian Diabetes Association to PCK, and an Alberta Innovates Health Solutions postdoctoral award to PCK, TP and BZ.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.