Animal Models of Dysregulated Cardiac Metabolism

Abstract

As a muscular pump that contracts incessantly throughout life, the heart must constantly generate cellular energy to support contractile function and fuel ionic pumps to maintain electrical homeostasis. Thus, mitochondrial metabolism of multiple metabolic substrates such as fatty acids, glucose, ketones and lactate are essential to ensuring an uninterrupted supply of ATP. Multiple metabolic pathways converge to maintain myocardial energy homeostasis. The regulation of these cardiac metabolic pathways has been intensely studied for many decades. Rapid adaptation of these pathways is essential for mediating the myocardial adaptation to stress, and dysregulation of these pathways contribute to myocardial pathophysiology as occurs in heart failure and in metabolic disorders such as diabetes. The regulation of these pathways reflects the complex interactions of cell-specific regulatory pathways, neurohumoral signals and changes in substrate availability in the circulation. Significant advances have been made in the ability to study metabolic regulation in the heart, and animal models have played an central role in contributing to this knowledge. This review will summarize metabolic pathways in the heart and describe their contribution to maintaining myocardial contractile function in health and disease. The review will summarize lessons learned from animal models with altered systemic metabolism, and those in which specific metabolic regulatory pathways have been genetically altered within the heart. The relationship between intrinsic and extrinsic regulators of cardiac metabolism and the pathophysiology of heart failure and how these have been informed by animal models will be discussed.

Introduction

To enable cardiac contraction, cardiac myocytes convert chemical energy derived from substrate utilization to mechanical work. Thus, the heart consumes more energy than other organs to meet the energetic needs of maintaining continuous pump function.¹ This inextricable link between energy metabolism and contractile function is reflected by the heart’s ability to turn over its entire ATP pool within a few seconds to match the constantly high energetic demand. Defects and changes in cardiac energy metabolism have been widely reported in subjects suffering heart failure (HF) with reduced ejection fraction (HFrEF), including impaired mitochondrial oxidative capacity, reduced cardiac efficiency, and alterations in the substrate utilization pattern.² The resulting energy deprivation is widely accepted to contribute to onset and progression of HF. In addition, alterations in systemic substrate availability determine changes in myocardial substrate utilization that are associated with altered cardiac structure and function, as exemplified in the metabolic syndrome (MetS).³ Thus, dysregulation of cardiac energy metabolism as a primary cause or secondary consequence may have detrimental consequences for cardiac structure and function.

Powerful non-invasive tools exist in clinical practice to evaluate cardiac metabolism, such as magnetic resonance spectroscopy (MRS) and positron emission tomography (PET). However, due to limitations in performing mechanistic experiments in human subjects and biomaterials, elucidating underlying mechanisms for altered cardiac energetics and development of therapeutic strategies to modulate myocardial energy metabolism remain challenging. Thus, investigators have generated and explored numerous animal models of dysregulated cardiac metabolism to facilitate mechanistic insights into the regulation of cardiac metabolism and to understand their relation to cardiac function and disease. Indeed, numerous findings from these models have been confirmed in human cardiac tissue thereby providing the rationale for clinical trials that partially reported beneficial clinical effects of targeted modulation of cardiac substrate utilization. Given the growing interest in developing metabolic therapies for the heart, we will present and discuss changes in cardiac metabolism in animal models of the metabolic syndrome and of HF and will review genetically engineered and disease-mimicking animal models useful to examine cardiac metabolism and related effects on cardiac function and structure. The large number of reported models precludes the inclusion of extensive details and citations in the main manuscript. Thus, we refer the reader to Supplemental Tables S1-S4 for detailed and comprehensive information on animal models and the relevant citations.

Overview of cardiac metabolism

Cardiac energy metabolism and its regulation has been the subject of intense investigation for several decades and has been reviewed in great detail by others.² In brief, about 95% of ATP generated in cardiac muscle originates from oxidative phosphorylation within mitochondria, and 5% of ATP is directly generated by glycolysis. To meet the high energy demands of the heart, cardiomyocytes possess a high density of mitochondria, comprising up to 40% cellular volume, depending on the species. Approximately 40-60% of ATP generated within mitochondria result from the oxidation of fatty acids (FAs), and the remaining from oxidation of glucose, lactate, ketone bodies and amino acids, based in part on their availability. Following uptake, predominantly by FA transporters (CD36, fatty acid transport protein (FATP)), FA are esterified to generate acyl-CoA by acyl-CoA synthetase, imported into mitochondria by the carnitine palmitoyltransferase (CPT) shuttle, and oxidized in the beta oxidation spiral to yield acetyl-CoA. Glucose is taken up via glucose transporters 1 (GLUT1; insulin-independent) and 4 (GLUT4; insulin-dependent), metabolized via glycolysis to generate pyruvate, which is imported into mitochondria via the mitochondrial pyruvate carrier and decarboxylated by pyruvate dehydrogenase (PDH) to also yield acetyl-CoA. Acetyl-CoA is oxidized in the TCA cycle, thereby mainly generating NADH₂ for delivery of electrons into the mitochondrial electron transport chain (ETC). Electrons are also delivered by FADH₂ generated by beta oxidation. Electrons passing through the ETC ultimately reduce oxygen to water, a process coupled to the regeneration of ATP from ADP by the F_oF₁-ATP synthase (ATPsyn). Other substrates enter this main pathway of substrate oxidation at different sites. Lactate can be imported into cardiomyocytes and is converted into pyruvate via lactate dehydrogenase (LDH) to enter substrate oxidation via PDH, accounting for < 3% of ATP in healthy hearts.⁴

Ketone bodies are predominantly generated in the liver, yielding acetoacetate which is converted to b-hydroxybutyrate (βOHB). βOHB is imported into mitochondria and converted to acetyl-CoA for oxidation in the TCA cycle by βOHB dehydrogenase (BDH1), succinyl-CoA:3-oxoacid-CoA transferase (SCOT), and mitochondrial thiolase, accounting for 6-7% of ATP.⁵ The major amino acid species oxidized by the heart are branched chain amino acids (BCAA, i.e., leucine, valine, isoleucine). BCAAs enter the cell via the BCAA:cation symporter family (LIVCS) and are converted to ketoacids by mitochondrial branched chain aminotransferase (BCATm). In a second reaction, branched chain alpha-keto acid dehydrogenase (BCKDH) converts these ketoacids to acetyl CoA and succinyl CoA which serve as oxidative or anaplerotic substrates, respectively.^{4, 6} The contribution to ATP synthesis is estimated at 4-5% of total ATP produced in healthy hearts.^{4, 6}

Regulation of oxidative capacity and metabolic substrate utilization is coordinated by a number of transcriptional, translational, and posttranslational mechanisms.^{2, 7, 8} The fetal heart predominantly relies on glycolysis for ATP production (accounting for up to 44% of total ATP), with low rates of oxidative ATP regeneration.⁹ Following birth, increased availability of oxygen and dietary FAs is accompanied by a 3-fold increase in mitochondrial density and a 10-fold increase in FA oxidation.^10-12 This increased mitochondrial content is driven by induction of the transcriptional coactivator and master regulator of mitochondrial biogenesis, peroxisome proliferator-activated receptor (PPAR)γ co-activator 1a (PGC1α), thereby increasing mitochondrial number, size and enzymatic content.¹² PGC1α coactivates and/or increases the expression of powerful transcription factors that drive FA oxidation and OXPHOS gene expression, including estrogen-related receptor (ERR) α and γ, and nuclear respiratory factor (NRF) 1 and 2, that increase mtDNA replication and transcription (mitochondrial transcription factor A, TFAm). Orchestration of mitochondrial biogenesis, the response to stress, adaptation to increased energy demands, regulation of OXPHOS function, maintenance of mitochondrial structure, and quality control require a balance of mitochondrial dynamics, a conserved process regulated by mitochondrial fusion (mitofusin 1/2 (MFN1/2), optic atrophy 1 (OPA1)) and fission proteins (dynamin-related protein 1 (DRP1), fission 1 (FIS1)). The predominant regulators of the expression of most enzymes involved in FA uptake and oxidation are the nuclear receptors and transcription factors, PPARα, γ, and β/δ.⁸ Another important regulator of cardiomyocyte energy homeostasis is the serine-threonine kinase, adenosine monophosphate-activated protein kinase (AMPK). AMPK is activated by a low AMP/ATP ratio (e.g. during exercise) and predominantly promotes catabolic metabolism by increasing glucose uptake via GLUT4 translocation to the sarcolemma, enhancing glycolysis by activating phospho-fructokinase-2 (PFK2), increasing FA uptake via CD36 translocation to the sarcolemma, and amplifying PGC1α signaling.¹³ These mechanisms are complemented by the classical glucose-FA cycle (also termed Randle cycle) wherein increased FA utilization leads to an inhibition of glucose utilization. Conversely, this competitive mechanism for fuel selection also applies when increased glucose utilization suppresses FA utilization (i.e., reverse Randle cycle). The molecular mechanisms for this regulation is detailed elsewhere.¹⁴ An important metabolite inhibiting CPT1β thereby limiting FA oxidation, is malonyl-CoA, which is generated by acetyl-CoA carboxylase (ACC) and catabolized by malonyl-CoA decarboxylase (MCD). These enzymes are also regulated by AMPK. Significant posttranslational regulation of energy metabolism is also mediated by the Sirtuin family of NAD⁺-dependent lysine deacetylases (Sirtuins 1-7 (SIRT1-7)), which are localized in the nucleus, cytosol and mitochondria and which remove a variety of modifications (e.g., acetyl-, succinyl-, malonyl, glutaryl-residues) to mainly activate their target enzymes and proteins. Figures 1 and 2 provides an overview of cardiac energy metabolism and mechanisms that regulate mitochondrial structure and function.

Figure 1A. Overview of energy substrate metabolism in the adult non-stressed heart.

1) Fatty acid (FA) uptake is facilitated by lipoprotein lipase (LPL) and mediated by FA transport proteins (FATP1/6) or CD36. FA are then esterified by acyl-CoA synthetase (ACS) to fatty acyl-CoA (FA-CoA), imported into mitochondria via the carnitine palmitoyltransferase (CPT) shuttle, and introduced for oxidation by medium-chain, long-chain or very long-chain acyl-CoA dehydrogenase (MCAD, LCAD, VLCAD). Activity of CPT1 is inhibited by malonyl CoA, which is generated by acetyl-CoA carboxylase (ACC) activity or decarboxylated by malonyl-CoA decarboxylase (MCD), an enzymatic system regulated by AMP-activated protein kinase (AMPK). Alternatively, FA-CoA may enter the cycle of synthesis and lipolysis of TG, regulated by diacylglycerol O-Acyltransferase (DGAT) and adipose triglyceride lipase (ATGL), or may be converted into other lipid intermediates (e.g., ceramides, DAG). 2) Glucose is taken up via glucose transporters (GLUT1/4), processed through glycolysis to yield pyruvate, which is imported into mitochondria by mitochondrial pyruvate carrier (MPC) and metabolized to acetyl-CoA by pyruvate dehydrogenase (PDH), the latter enzyme being inhibited by PDH kinase 4 (PDK4) activity. Alternatively, early glycolytic intermediates can be diverted into glycogen synthesis or into the pentose phosphate pathway (PPP) or hexosamine pathway (HP). Glucose uptake is increased by insulin signaling via insulin receptor (IR) and insulin-like growth factor 1 receptor (IGF1-R), thereby increasing GLUT4 translocation to the sarcolemma. Lactate is taken up by monocarboxylic acid transporter (MCT) and converted to pyruvate by lactate dehydrogenase (LDH). 3) Following uptake by monocarboxylate transporter (MCT) 1, ketones (predominantly β-hydroxybutyrate, βOHB) are converted to acetoacetate (AcAc) by βOHB dehydrogenase 1 (BDH1) and further oxidized to acetyl-CoA by succinyl-CoA:3 oxoacid-CoA transferase (SCOT) and thiolase. 4) Branched chain amino acids (BCAA) are imported into mitochondria by the solute carrier family 25 member 44 (SLC25A44), converted to branched chain ketoacids (BCKA) by mitochondrial branched chain aminotransferase (BCATm), and then to acetyl-CoA by branched chain alpha-keto acid dehydrogenase (BCKDH). BCKDH activity is inhibited by BCKD kinase and activated by mitochondrial protein phosphatase 2C (PP2Cm). 5) Acetyl CoA yielded from FA, carbohydrate, ketone or BCAA utilization enter the tricarboxylic acid (TCA) cycle to produce NADH₂. NADH₂, and FADH₂ generated in β-oxidation, deliver electrons into the electron transport chain (ETC) to sustain oxidative phosphorylation (OXPHOS). The resulting ATP maintains cardiac contractility and fuels ionic pumps.

Figure 2. Mitochondrial life cycle.

Peroxisome proliferator-activated receptor gamma coactivators 1-α/β (PGC1α/β) coactivate peroxisome proliferator-activated receptors (PPARs), estrogen-related receptors (ERRs) and nuclear respiratory factors (NRFs) to orchestrate expression of proteins of FA utilization and OXPHOS during mitochondrial biogenesis. PGCs also drive mitochondrial transcription factor A (TFAm) expression to regulate protein/RNA expression and replication of mtDNA. Mitochondrial dynamics are balanced by mitochondrial fission, driven by dynamin-related protein 1 (DRP1) and fission 1 (FIS1), and mitochondrial fusion, promoted by mitofusin 1/2 (MFN1/2) and optic atrophy 1 (OPA1). Removal of mitochondria is accomplished by autophagy of mitochondria (i.e., mitophagy).

Relationship between cardiac metabolism and function

A bidirectional dependency exists between cardiac metabolism and function where changes in contractility demand adaptations in energy metabolism, and where changes in cardiac metabolism alter contractile function. Following an acute increase in workload (e.g., during acute exercise or increased blood pressure), the heart may increase its work by up to 3-fold, and in trained athletes by up to 6-fold. This contractile reserve requires an increase in coronary blood flow to match delivery of oxygen and substrates to increased energy demand.¹⁵ Simultaneously, the heart can increase substrate flux and oxidative phosphorylation, which operates at only 15-25% of its maximal capacity under resting conditions.¹⁶ However, under maximal workload, the heart may utilize up to 90% of its ATP-generating capacity within mitochondria. In addition, the heart utilizes cardiac metabolic reserve present in the form of phosphocreatine, TG and glycogen. Phosphocreatine serves as a rapidly claimable source of phosphate to regenerate ATP, and enzymatic breakdown of glycogen can rapidly feed glucose-6-phosphate into glycolysis for glucose oxidation.¹⁷

The general dependency of contractility on ATP availability is supported by studies using animal models with selected genetic manipulations. For example, mice lacking PGC1α have diminished content and activity of mitochondrial energy-metabolic proteins, resulting in mildly impaired contractile function in ex vivo beating Langendorff-perfused hearts, but clearly impaired ability to increase work output in response to dobutamine stress.¹⁸ These mice revealed a 20% reduction of ATP content under non-stressed conditions and impaired ability to increase ATP turnover that likely limited cardiac contractile performance in response to dobutamine stress. These observations provide evidence that a primary dysregulation of energetics (induced by PGC1α deficiency) is sufficient to impair cardiac contractility and reserve. Accordingly, reduced expression and signaling of PGC1α and the related decrease in OXPHOS expression in pathological hypertrophy and failing hearts have been proposed to contribute to onset or progression of heart failure.^19-22 Although defects in cardiac energetics can induce or exacerbate contractile dysfunction, changes in cardiac structure and function are also correlated with and may be mechanistically linked to changes in metabolism.

Cardiac metabolism and function are also impacted by changes in systemic metabolism. As an omnivore the heart oxidizes a variety of substrates dependent on their availability in the circulation. This flexibility allows the heart to sustain pump function even under metabolically unfavorable conditions to mitigate energy deficiency. This ability to switch substrate preference for ATP generation (e.g., from FA utilization to glucose utilization) is mediated by acute and chronic signaling within cardiomyocytes that adapts the activity and capacity of metabolic enzymes to accomplish the shift in substrate utilization, a concept described as metabolic flexibility. For example, the heart increases the relative utilization of glucose during fed conditions when serum glucose and insulin levels are high, whereas starvation induces lipolysis leading to increased serum FA levels and cardiac FA utilization. Furthermore, high serum levels of FAs and triglycerides in obesity and diabetes mellitus cause a predominant use of FAs as the primary source for ATP regeneration. Loss of metabolic flexibility, as occurs in heart failure or the cardiomyopathy of obesity and diabetes mellitus, contributes to decreased cardiac efficiency and myocardial energy deficiency.

Overview of the metabolic syndrome and impact on cardiac metabolism, structure and function.

Recent decades, have been characterized by a global obesity epidemic, accompanied by increased incidence of the metabolic syndrome (MetS).²³ This is characterized by the presence of insulin resistance (i.e., impaired fasting glucose, impaired glucose tolerance, or type 2 diabetes mellitus (T2DM)), combined with at least two other risk factors, including obesity, dyslipidemia (hypertriglyceridemia, low high-density lipoprotein cholesterol), or hypertension.²⁴ While the pathogenesis of the MetS is complex and incompletely understood, genetic predisposition and the interaction with environmental factors, and sedentary lifestyle (lack of exercise, excess caloric intake) leading to visceral adiposity are highly associated with the MetS.²⁵ Insulin resistance in adipose tissue increases lipolysis and circulating FAs that may impair glucose uptake and induce insulin resistance in skeletal muscle, induce hepatic insulin resistance and thus lipogenesis and gluconeogenesis and also impair pancreatic beta cells function resulting in decreased insulin secretion.^{25, 26} While insulin resistance may initially be compensated by hyperinsulinemia, reduced beta cell reserve ultimately leads to hyperglycemia and type 2 diabetes in susceptible individuals. Dysregulation of adipokine secretion (decreased adiponectin and increased leptin release), activation of the renin-angiotensin-aldosterone system (RAAS) reflected by increased secretion of angiotensin II, and release of proinflammatory cytokines (e.g. TNF-α, IL-1β, IL-6) contribute to insulin resistance in peripheral tissues by direct and indirect mechanisms such as oxidative stress, fibrosis, metabolic derangements, and inflammation.^{26, 27} Loss of the vasodilator effect of insulin, vasoconstriction caused by FFAs, increased RAAS and sympathetic nervous system activation, and induction of endothelial dysfunction may contribute to development of hypertension in the MetS.²⁸

While increasing the risk for atherosclerotic cardiovascular disease (CVD), the MetS also promotes the development of a cardiomyopathy of obesity and diabetes. Some individuals may develop concentric hypertrophy and diastolic dysfunction, possibly progressing to heart failure with preserved ejection fraction (HFpEF), whereas others may develop eccentric hypertrophy, occasionally accompanied by systolic defects in contractility and potentially progressing to heart failure with reduced ejection fraction (HFrEF).^{29, 30} Numerous studies using rodent models of obesity, IR and T2DM provide evidence that these metabolic disorders may promote the development of cardiac hypertrophy with increased cardiomyocyte size, myocardial fibrosis, and functional defects such as impaired contractility, diastolic dysfunction and impaired cardiac efficiency (i.e. cardiac work/oxygen consumption).^31-36 It is important to note that these models are generally resistant to the development of atherosclerosis in the absence of predisposing genetic mutations (e.g., ApoE or LDLR deficiency), thus emphasizing that metabolic perturbations per se could be sufficient to induce cardiac dysfunction. As summarized in Figure 3, studies investigating cardiac metabolism in these models revealed increased uptake and oxidation of long chain FAs, decreased glucose utilization, and diminished metabolic responsiveness upon insulin stimulation compatible with cardiac insulin resistance.^{33, 37, 38} In addition, impaired rates of mitochondrial ATP synthesis have been reported in hearts of various animal models of metabolic disease.^{32, 37} Underlying mechanisms include transcriptional, translational and posttranslational repression of OXPHOS activity, mitochondrial oxidative stress, mitochondrial uncoupling, dysregulation of mitochondrial biogenesis, perturbation of mitochondrial dynamics, and altered mitophagy, among others.^{19, 32, 39, 40} Of note, some changes described in rodent models have been confirmed in hearts of obese, insulin resistant and/or type 2 diabetic subjects, including increased FA utilization, decreased cardiac efficiency, increased mitochondrial ROS generation, and defects in mitochondrial respiratory function.^{30, 41-44}

Figure 3: Simplified overview of cardiac energy metabolism in the Metabolic Syndrome.

Increased delivery of FA and TG, as well as activation of PPARα signaling, increase FA oxidation and generation of reactive lipid intermediates (acyl-CoA, ceramide, DAG). Glycolysis and glucose oxidation (GO) are decreased, contributed to by decreased glucose uptake due to impaired insulin stimulated glucose uptake. Early glycolysis intermediates are shifted into the PPP, AGE generation, and HBP pathway, driving O-Glc-Nacylation that could contribute to cardiac remodeling. KB oxidation may be increased, whereas BCAA oxidation is decreased, contributing to PDH inhibition and cardiac hypertrophy by mTOR activation. Mitochondrial ATP synthesis is impaired due to defects in the respiratory chain and ROS-mediated mitochondrial uncoupling. Mitochondrial and cytosolic ROS impair Ca²⁺ handling, induce fibrosis, promote inflammation and increase cell death. Collectively, impaired energetics contribute to cardiac remodeling, dysfunction and inefficiency.

Animal models of the metabolic syndrome

A variety of animal models exhibit dysregulated cardiac metabolism resulting from systemic changes associated with the MetS (summarized with detailed phenotypes in Supplemental Table S1). Specific mechanisms by which perturbed whole-body metabolism and changes in circulating levels of glucose, FA, cholesterol and insulin, influence cardiac energy metabolism, structure and function varies considerably between animal models.

Role of species differences

The murine and human genomes are similar in size, number of genes (>99%) and conservation of gene order.⁴⁵ Since mouse strains are mostly inbred, genetic differences among these strains and also their sub-strains cause marked differences in systemic biochemical and metabolic phenotypes. These changes in turn may contribute to differences in the metabolic, structural and functional responses of the heart.^{46, 47} These inter-strain genetic differences have been exploited to explore the contribution of specific gene products to phenotypic abnormalities, including systemic dysmetabolism and their effects on cardiac function and structure.^{47, 48} Additional considerations such as choice of controls and effects of mixed backgrounds have been reviewed in detail elsewhere.⁴⁷ Due to their size, measurement of some metabolic and cardiac parameters may be easier in rat models of MetS or cardiac disease. For both mice and rats, the age of animals and specific treatment regimens (e.g., type and duration of a diet) need to be considered among different rodent species. Thus, careful selection of the rodent model is advised depending on the research questions to be addressed, and care should be taken when extrapolating findings from rodent studies to humans with large genetic heterogeneity.

High-fat diet

High-fat diets utilized in rodent models primarily range from 45-60% of total energy (kcal) from fat sources, mainly from saturated FAs. In rats, short-term (3 weeks) feeding with high-fat diet (HFD) increases plasma FA and cholesterol levels, accompanied by obesity. Hearts display increased myocardial O₂ consumption (MVO₂) without an increase in cardiac function, resulting in decreased cardiac efficiency, likely related to FA-induced uncoupling of ATP synthesis from O₂ consumption.³⁷ Longer HFD feeding (7-8 weeks) induces systemic insulin resistance, accompanied by cardiac hypertrophy, and diastolic and mild systolic dysfunction; however, most HFDs fail to induce profound alterations in cardiac remodeling and function in rats. Increased FA uptake by relocation of CD36 to the plasma membrane and concurrent impairment in mitochondrial FA import and oxidation may result in FAs being shifted away from oxidation to storage.⁴⁹ One study demonstrated increased FAO in response to 48 weeks of HFD (60%), whereas a Western diet with less fat (45%) initially increased FAO but finally decreased oxidation rates, accompanied by cardiac dysfunction.⁵⁰ Selective cardiac dysfunction induced by Western diet may be related to inadequate stimulation of PPARα-driven expression of FAO genes, or inadequately increased FA-induced mitochondrial uncoupling.

In mouse models, 8-12 weeks of HFD increases body weight and circulating glucose, FA, cholesterol, TG and insulin levels.⁵¹ Similar to rats, minimal consequences on LV hypertrophy/remodeling and function are observed. While exhibiting enhanced FAO and reduced glucose oxidation, sustained mitochondrial function suggests cardiac energy status to be unaffected. Inhibition of PDH indicative of metabolic inflexibility may develop as early as day 1 of a HFD, although development of metabolic insulin resistance of the heart may require several weeks of HFD.⁵² Along with cardiac TG accumulation, increased content of reactive long-chain acyl-CoAs, ceramide, cholesterol, DAG and acylcarnitines is also observed, suggesting the presence of cardiac lipotoxicity in these models. With longer duration of HFD (> 20 weeks), myocardial levels of OXPHOS proteins may also decrease but do not seem to impair respiratory function or induce energy depletion, and ketogenic enzymes appear to be affected, although whether they are induced or repressed appears to be conflicting.⁵³ In some studies cardiac hypertrophy and dysfunction, systolic dysfunction is absent following long-term HFD,⁵⁴ although in other studies LV dysfunction has been reported.⁵⁵ The basis for these disparities is not understood. Additional mechanisms such as increased oxidative stress, apoptosis, fibrosis and impaired Ca²⁺-handling may contribute to cardiac remodeling and functional defects in these models.

High-sugar or High-fat/high-sugar diet

Unlike HFD, high-sugar diets (high-sucrose or high-fructose) do not appear to induce obesity in rats in the first 9-12 weeks, but only beyond 30 weeks.⁵⁶ However, hyperlipidemia and systemic insulin resistance are observed early on. High-sugar diets induce systolic and diastolic dysfunction in some studies, although the development of cardiac hypertrophy and myocardial fibrosis appears to be less obvious.⁵⁶ Reduced cardiac ATP/ADP ratio and PCr content suggests an energetic deficit, possibly caused by reduced expression of genes involved in FA transport and oxidation. An association with increased cardiac lyso-phosphatidyl choline and FA content was also observed, which may induce cardiomyopathy by increasing apoptosis.

Increasing the fat content of the high sugar diet in mice, is insufficient to induce an overt cardiac phenotype following short-term feeding (1-5 weeks). However, longer (> 4 months) duration may result in the development of cardiac hypertrophy, dysfunction, and impaired contractile reserve.⁵⁷ This diet induces a pattern of substrate utilization that mimics changes observed in longstanding obesity or T2DM, characterized by an early reduction of cardiac glycolysis and glucose oxidation rates (likely mediated by decreased GLUT4 content and translocation), increased FAO rates, and blunted cardiac metabolic response to insulin, accompanied by impaired ATP regeneration indicative of myocardial energy deprivation.⁵⁸ The latter may be related to oxidative damage of energy metabolic proteins involved in FA utilization, glycolysis, TCA cycle or OXPHOS. As in HFD models, the early decrease in glucose utilization may be the initial molecular defect leading to a shift in the pattern of myocardial substrate oxidation.

Chemical-induced models of diabetes and MetS

Despite potential toxic side effects, chemical-induced diabetes may be a convenient model when exploring effects in rodents with genetic manipulations. Treatment with the beta-cell toxin, streptozotocin (STZ), leads to insulin-deficient diabetes with hyperglycemia and hyperlipidemia but not obesity. Cardiac size is reduced in STZ-treated mice, likely due to the increased protein degradation resulting from insulinopenia, and cardiac contractility is often impaired.⁵⁹ In most studies, hearts typically display reduced glucose uptake, increased FA uptake and oxidation, and increased incorporation of FA into cardiac phospholipids, TG and FA.^{60, 61} When FA utilization is increased, MVO₂ tends to increase as well, leading to reduced cardiac efficiency.³⁶ STZ-induced diabetes increases circulating ketone body (KB) levels, particularly the ratio of βOHB to acetoacetate. While it remains unclear whether cardiac βOHB uptake is increased or decreased in STZ-induced diabetes, reduced KB utilization observed in some studies may be related to decreased expression and activity of BDH1, or to tyrosine nitration of SCOT.⁵³ Of note, STZ treatment has also been combined with HFD in few studies to model severe, uncontrolled type 2 diabetes and to recapitulate additional traits (e.g. obesity, insulin resistance) of that characterize human type 2 diabetes in this model.

Genetic models of MetS

Several genetic animal models have provided useful insight to understanding the interaction between perturbed systemic metabolism, altered cardiac substrate utilization and development of cardiac dysfunction. Ob/ob mice carry a recessive mutation in the leptin gene resulting in deficiency of a functionally active leptin protein. Compared to their lean counterparts, ob/ob mice develop obesity and systemic insulin resistance as early as 4 weeks of age, with progressive increases in FA, TG and insulin levels as they age. The cardiac phenotype includes development of cardiac hypertrophy, variable degrees of systolic dysfunction, increased FAO and reduced glucose oxidation as early as 4 weeks of age.³⁴ Studies using Langendorff heart perfusions demonstrated that the presence of palmitate in the perfusion buffer (but not glucose alone) resulted in reduced cardiac efficiency, associated with impaired ATP synthesis and less ATP generated per oxygen consumed when subsequently measured in saponin-permeabilized myocardial fibers, suggesting that FA-induced mitochondrial uncoupling may have reduced cardiac efficiency.³⁷ In addition, ob/ob hearts display metabolic insulin resistance, evidenced by the inability of insulin to increase cardiac glucose utilization, likely due to impaired AKT signaling and subsequent GLUT4 translocation, and to suppress FA oxidation.³³ Altogether, studies in ob/ob hearts revealed impaired metabolic flexibility and impaired conversion of ATP into mechanical work as a central characteristic of obesity and diabetes-associated cardiomyopathy .

Db/db mice carry a leptin receptor defect and on a susceptible genetic background, are characterized by obesity, hyperglycemia, increased circulating FA, TG, cholesterol and insulin, and impaired glucose tolerance as early as 8 weeks of age. In most studies, hearts were reported to be hypertrophied with some degree of systolic and diastolic dysfunction.³⁶ Similar to ob/ob mice, myocardial FA uptake, FAO and also TG accumulation are increased, whereas basal glycolysis, glucose oxidation, and insulin-stimulated glucose utilization are impaired. FA-induced mitochondrial uncoupling and impaired cardiac efficiency was also confirmed in db/db hearts.³² Furthermore, db/db hearts displayed impaired ATP regeneration, in part from reduced expression and posttranslational modification (acetylation, deamidation, oxidation) of OXPHOS subunits thereby supporting the concept that mitochondria are important targets of diabetes-induced cardiac injury.

Another commonly used genetic rodent model of obesity is the Zucker fatty (ZF) rat which carries a non-functioning leptin receptor secondary to a missense mutation in the leptin receptor gene. These rats are characterized by obesity and increased serum levels of FA, TG and insulin. Selective breeding of hyperglycemic ZF rats yielded the inbred sub-strain of Zucker diabetic fatty rats (ZDF rats), which show early insulin resistance and later (10-12 weeks) develop, hypoinsulinemia and hyperglycemia in contrast to ZF rats. ZF rats develop cardiac hypertrophy, whereas cardiac function was reported to be either reduced or increased.⁶² Increased FA uptake is associated with increased TG storage, FA oxidation is not to increased, which may explain the absence of impaired cardiac efficiency in this model. In ZDF rats, cardiac hypertrophy is more consistently accompanied by contractile defects, and rates of FA oxidation are increased, whereas glucose oxidation was lower.⁶³ ZDF rats have also been crossed with spontaneously-hypertensive (SHR) rats (ZSF1), a model with similar systemic metabolic abnormalities as observed in ZDF rats with additional hypertension that develop cardiac hypertrophy and diastolic dysfunction. While data on actual metabolic rates in hearts of ZSF1 rats are lacking, pathway analysis of the cardiac transcriptome revealed enrichment of genes encoding for enzymes of FA and BCAA utilization, including induction of CPT-1 and PDK4, suggesting increased FAO and decreased glucose oxidation as observed in ZDF rats.⁶⁴

When comparing mouse and rat models of impaired leptin action, the ZF rat is the only model that does not develop hyperglycemia, thus allowing analysis of the long-term effects of obesity without confounding effects of hyperglycemia. Relative to ob/ob and db/db mice, serum lipid levels appear to be altered more dramatically in the ZDF rat. Given that ZF rats are an outbred strain with genetic heterogeneity, the model may more closely resemble the human disease compared to inbred mouse and rat strains. Of note, in all mouse and rat models with leptin deficiency or leptin receptor dysfunction, specific effects on the systemic and cardiac phenotype that could be a direct consequence of impaired leptin action cannot be excluded.⁶⁵

Another genetic model of elevated blood glucose is the OVE26 mouse, in which overexpression of calmodulin in pancreatic beta cells causes cellular damage, insulin-deficient diabetes and hypertriglyceridemia. While data on substrate oxidation rates are lacking, unchanged levels of GLUT4 and insulin-stimulated AKT phosphorylation but increased phosphorylation of the PDH subunit E1 and impaired rates of pyruvate-supported respiration suggest intact myocardial insulin signaling but impaired glucose oxidation in this model.⁶⁶ In fact, impaired mitochondrial respiration with glutamate or succinate indicates a general oxidative defect, likely related to oxidative stress. While OVE26 mice display cardiac hypertrophy without functional impairment at 14 weeks of age, dilated cardiomyopathy becomes apparent at 6 months of age.

The βV59M mouse develops insulin deficiency due to beta-cell specific expression of a human activating K_ATP channel mutation. Of note, insulin deficiency and hyperglycemia develop 24h after transgene induction, whereas hyperlipidemia does not develop until 4 weeks following transgene induction.⁶⁷ In the absence of hyperlipidemia, hearts of βV59M mice have decreased cardiac output and show decreased PDH flux, while FA content was decreased and content of glycogen or lipid was unchanged, suggesting ATP production preferentially from FAO and uncoupling of glycolysis from glucose oxidation. Proteomics revealed increased expression of PDK4 and of proteins involved in FAO and lipolysis, while expression of LDH, and glycolytic and TCA cycles enzymes were reduced. Thus, compared to other models, βV59M mice develop changes in cardiac function and metabolism that are at least in part independent of increased circulating FA. Akita mice develop insulin-deficient diabetes secondary to a defect in insulin processing. Hearts of these mice exhibit reduced glucose utilization, increased FAO myocellular lipid accumulation and mitochondria dysfunction, despite preserved insulin sensitivity.^{68, 69}

Of note, models have been generated that combine metabolic traits of MetS with risk for spontaneous atherosclerosis, to more closely parallel human MetS. Recently, ApoE-KO mice were crossed with hypertensive BPH/2J Schlager mice (BPHxApoE-KO), which display elevated blood pressure, impaired glucose tolerance, and increased serum levels of cholesterol, TG, FA and insulin upon feeding a Western diet.⁷⁰ Data on cardiac function and energy metabolism remain to be reported.

Lessons learned from genetic models of altered cardiac metabolism or metabolic signaling

Numerous models with global or cardiomyocyte-restricted genetic manipulations of energy metabolic proteins or their regulators have been generated to mimic specific traits of metabolic disease and CVD, or to allow insights into their role in metabolic and cardiovascular pathologies. These models will be briefly discussed below, sorted by their attribution to specific biological pathways (Table 1). Detailed phenotypes can be found in Supplemental Table S2. While genetic models provide valuable insights into pathophysiologic mechanisms, it remains a challenge to relate a genetic model (where a protein/enzyme has been completely deleted) to a pathological condition relevant to humans. Thus, studying animal models of human disease, as opposed to diseased animal models, would be generally preferable. At this point, we would also like to remind the reader that transcriptional changes do not always correlate with protein levels and/or protein function or enzyme activity. This is particularly true in energy metabolism and requires careful interpretation of data provided in tables and throughout the manuscript. We also refer the reader to a scientific statement from the American Heart Association elaborating on methods and models that can be used to investigate cardiac metabolism.⁷¹ Finally, this review focuses mainly on metabolic pathways in cardiomyocytes. Insights in non-myocytes such as endothelial cells⁷² and fibroblasts as recently reviewed⁷³ are now emerging and will likely contribute additional perspectives in the future.

Table 1:

Overview of genetically engineered mouse models mimicking or exhibiting dysregulated cardiac metabolism.

Metabolic Pathway	Loss-of-function	Gain-of-function
FA uptake	CD36-KO, cmCD36-KO (+ aging); ACSL1-KO, cmACSL1-KO	CD36-OE (in SHR rats), cmFATP1-OE, ACSL1-OE, cmACSL1-OE
Triglyceride synthesis and lipolysis	cmDGAT1-KO; ATGL-KO (+ STZ), cmATGL-KO; LPL-KO, cmLPL-KO	cmDGAT1-OE; cmATGL-OE (+ HFD/high-sugar diet or STZ); LpL-OE, cmLpL-OE
Mitochondrial FA transport	ACC2-KO, ACC2-KI (mutant), cmACC2-KO; MCD-KO (+ HFD or aging); CPT1b-KO	L-CPT1-OE
FA oxidation:	VLCAD-KO, cmVLCAD-KO; LCAD-KO
Regulators of energy metabolism	PGC1α-KO (+ HFD or in ob/ob mice); PGC1β-KO; PGC1α/β-KO (in ob/ob mice), cPGC1α/β-KO; TFAm-KO, cmTFAm-KO; POLGA-KO, cmMTERF3-KO; cmTFB1M-KO; NRF1-KO, efNRF2-KO; ERRα-KO, ERRγ-KO, cm ERRγ-KO, ERRα/cmERRγ-KO; PPARα-KO (+ STZ or HFD or in UCP-DTA mice); cmPPARδ-KO; PPARγ-KO, cmPPARγ-KO; AMPKα2-KO (+ calorie restriction); AMPKα2-OE (mutant), AMPKγ2-OE (mutant: N488I)	cmPGC1α-OE; TFAm-OE; cERRγ-OE; cmERRγ-OE; cmPPARα-OE (+ HFD or STZ); cmPPARγ-OE
Glucose utilization	PDK4-KO (+ HFD); cmPDK4-KO; cmGLUT1-KO, GLUT4-KO, cmGLUT4-KO (+ exercise), cmMPC1-KO, cm-MPC2-KO	cPDK4-OE; cmGLUT1-OE (+ HFD), GLUT4-OE (in db/db mice), cmGLUT4-OE (+ STZ); cPFK2-OE
Insulin/IGF1 signaling	CIRKO (+ exercise or STZ), cIRS1/2-KO, cIRS1-KO (+ exercise), cIRS2-KO (+ exercise); cmPI3K-KO; AKT1-KO (+ exercise), AKT2-KO (+ exercise), cmAKT-KO; cmIGF1R-KO (+ aging or + exercise); CIGF/IR-KO	cPI3K-OE; cmAKT-OE; IGF1R-OE; IGF-IIRα-OE
Ketone body utilization	cmBDH1-KO; SCOT-KO, cmSCOT-KO	cmBDH1-OE
BCAA utilization	cmBCATm-KO; KLF15-KO; PP2Cm-KO	SLC25A44-OE
Posttranslational regulation	SIRT1-KO (+ exercise), SIRT1-KO (+ HFD), SIRT3-KO (+ STZ or HFD/DOCA or HFD/L-NAME), ecSIRT3-KO, CrAT/SIRT3-KO; SIRT4-KO; SIRT5-KO, cmSIRT5-KO; SIRT6-KO, cmSIRT6-KO; SIRT7-KO, cmSIRT7-KO; cmOGT-KO, cmOGA-KO	SIRT1-OE, SIRT1/PPARα-OE, SIRT3-OE (in db/db mice); SIRT4-OE; SIRT6-OE (+ HFHS diet), cmSIRT6-OE
Mitochondrial dynamics	MFN1/2-KO, cmMFN1/2-KO, cmMFN1-KO, cmMFN2-KO; cmDRP1-KO; OPA1-KO; cmTME1L-KO; DRP1/MFN1/2-KO	cmDRP1-OE

See extended version and references in Supplemental Table S2.

Abbreviations: ACC, acetyl-CoA carboxylase; AKT, protein kinase B; AMPK, adenosine monophosphate-activated protein kinase; ACSL, acyl-CoA synthetase, long chain; BCAA, branched chain amino acid; BCATm, branched-chain aminotransferase, mitochondrial; BDH, β-hydroxybutyrate dehydrogenase; CD36, fatty acid transporter; cm, cardiomyocyte; CIRKO – cardiomyocyte KO of insulin receptor, cIRS(1/2)1KO cardiomyocyte KO of insulin receptor substrates (1/2); CPT, carnitine palmitoyltransferase; CrAT, carnitine palmitoyl transferase; DGAT, diacylglycerol acyltransferases; DOCA, deoxycorticosterone acetate; DRP, dynamin-related protein; EF, embryonic fibroblast; ERR, estrogen-related receptor ; FA, fatty acid; GLUT, glucose transporter; IGF, insulin-like growth factor; IGF1R, IGF-1 receptor; IR, insulin receptor; KI, knock-in; KLF15, Krüppel-like factor 15 ; KO, knockout; L-CPT, CPT liver isoform; LCAD, long-chain acyl CoA dehydrogenase; MCAD, medium-chain acyl-CoA dehydrogenase; MCD, malonyl-CoA decarboxylase; MFN, mitofusin; NRF, nuclear respiratory factor; MPC, mitochondrial pyruvate carrier; OE, overexpression; OGA, O-GlcNAcase; OGT, O-GlcNAc Transferase; OPA, optic atrophy; PDK, pyruvate dehydrogenase kinase; PFK, phosphofructokinase; PGC,1 PPARγ co-activator 1; PI3K, phosphoinositide 3-kinase; POLGA, DNA polymerase gamma, subunit A; PP2Cm, protein phosphatase 2C , mitochondrial; PPAR, peroxisome proliferator-activated receptor; SCOT, succinyl-CoA:3-oxoacid-CoA transferase; SIRT, sirtuin; SLC, solute carrier family; STZ, streptozotocin; TFAm, mitochondrial transcription factor A; TFB1M, transcription factor B1; UCP, uncoupled protein; VLCAD, very long-chain acyl-CoA dehydrogenase; YME1L, ATP-dependent metalloprotease

FA uptake and activation:

Several models have been engineered to explore the mechanisms of cardiac FA uptake and activation. Mice with CD36 deficiency revealed that CD36 mediates a major fraction of the FA uptake by myocardial tissue and is required to maintain oxidative ATP regeneration and cardiac function in response to chronic pressure overload.^{74, 75} However, in the aged heart, increased CD36 levels may promote myocardial lipid accumulation and cardiac dysfunction.⁷⁶ Studies in mice with combined CD36 deficiency and PPARα overexpression showed that CD36-mediated FA uptake is necessary for PPARα-driven lipotoxicity, and knockdown of CD36 was shown to protect from HFD-induced lipid accumulation, oxidative stress and cardiac dysfunction, collectively suggesting that inhibiting CD36-mediated FA uptake may be a promising strategy to prevent FA-induced cardiac dysfunction observed in obesity and diabetes.^{77, 78} Significance of FATP for FA uptake and oxidation was underscored by studies in mice with cardiomyocyte-selective overexpression of FATP, which consequently develop diastolic dysfunction.⁷⁹ Studies in mice with deficiency or overexpression of ACSL1 provided evidence that ACSL1 also regulates FA uptake but additionally activates FAs for mitochondrial oxidation by intracellular trapping of acyl units via esterification to CoA.^{10, 80-82} Since increased FA uptake by ACSL1 overexpression in cardiomyocytes dose-dependently induced hypertrophy or dilated cardiomyopathy, this model evolved as a convenient model to study consequences of cardiac lipotoxicity, including ROS-driven perturbation of mitochondrial dynamics as a novel mechanism of lipotoxicity.¹⁰ Recent analysis of low-level ACSL1 overexpression suggested that these mitochondrial adaptations could redirect long-chain Acyl CoAs into cardioprotective ceramides, which limited LV remodeling following pressure overload.⁸³

Mitochondrial FA import and oxidation:

Models with deficiency of FAO enzymes provided insights into underlying mechanisms of cardiomyopathy observed in some humans with inherited disorders of FAO. Studies in VLCAD^−/− and LCAD^−/− mice suggest energy deficiency due to impaired FAO and TCA cycle anaplerosis, and/or lipotoxicity due to accumulation of lipid intermediates such as ceramides as underlying mechanisms of associated cardiomyopathy.^{84, 85} Studies in mice with increased activity of LPL on the surface of cardiomyocytes revealed that LPL can increase uptake and storage of lipids derived from circulating TG, which may lead to cardiomyopathy.⁸⁶ Conversely, cardiomyocyte-restricted knockout of LPL was necessary to maintain cardiac FAO and contractile function, collectively indicating a physiological role of LPL in regulating cardiac energy metabolism and function.⁸⁷

Allosteric inhibition of CPT1B by malonyl-CoA is also a potent mechanism to inhibit FAO in the heart, as demonstrated in a knock in mouse model expressing the CPT1BE3A mutant enzyme, which has reduced sensitivity to malonyl-CoA.⁸⁸ While ACC2 catalyzes the conversion of acetyl-CoA to malonyl-CoA, the reverse reaction is driven by MCD. Models with deletion of ACC or MCD have provided valuable insights into cardiac metabolism. Mice with ACC2 deletion confirmed that decreasing malonyl-CoA levels disinhibits CPT1B and thereby increases myocardial FAO. Of note, this metabolic intervention resulted in a correction of the typical switch towards increased glucose utilization observed following TAC and attenuated cardiac hypertrophy, suggesting that the reversion to the fetal metabolic profile observed in pathological hypertrophy is not necessarily beneficial.⁸⁹ Of note, while AMPK-mediated phosphorylation and inhibition of ACC has been proposed to increase FAO, studies in mice with mutated AMPK phosphorylation sites on ACC1 and ACC2 showing no effect of FAO rates suggest that AMPK-dependent inactivation of ACC may not be essential for the control of myocardial FAO and subsequent cardiac function during conditions of stress known to activate AMPK such as increased workload or myocardial ischemia.⁹⁰ Mice with MCD deletion revealed increased glucose oxidation and cardioprotection following ischemia, suggesting chronic pharmacological MCD inhibition as a therapeutic strategy for cardioprotection.⁹¹ In diet-induced obesity, intracellular accumulation of lipid intermediates is thought to contribute to insulin resistance in skeletal and cardiac muscle. However, using mice with germline deletion of MCD paradoxically revealed that despite reduced induction of FAO and accumulation of FA intermediates insulin-stimulated glucose oxidation was not impaired in response to diet-induced obesity.⁹² These findings imply that myocardial accumulation of lipid intermediates might not be sufficient to impair glucose utilization and that inhibition of FAO by MCD inhibition may instead have therapeutic potential in preventing obesity-induced insulin resistance in the heart. Of note, inhibition of FAO in mice with CPT1B deficiency leads to accumulation of TG and ceramides in the heart, and this lipotoxicity may cause aggravation of cardiac hypertrophy and dysfunction of these mice in response to pressure overload.⁹³ Thus, underlying cardiac disease mode of FAO inhibition and potentially differences in lipid species that accumulate, may determine whether beneficial or detrimental effects will be achieved by interrogating FA utilization. Expression of the liver isoform of CPT1 (CPT1a) is induced in the failing heart and adenoviral overexpression of CPT1a in rat hearts reduced rates of FAO, indicating that reduced FAO in heart failure might be due in part to CPT isoform switching.⁹⁴

Triglyceride synthesis and lipolysis:

Fluctuations in myocardial TG content reflect or induce changes in cardiac energy metabolism. Simultaneous cardiomyocyte-selective deletion of diacylglycerol acyl transferase 1 and 2 (DGAT1 and 2) suppressed TG turnover and synthesis, demonstrating the significance of DGATs in the synthesis of cardiac TG.⁹⁵ In mice with increased FA uptake due to ACSL1 overexpression, simultaneous overexpression of DGAT1 lead to increased TG storage, lowering of toxic lipid intermediates and improvement of cardiac function, indicating a protective role of TG in states of lipid overload.⁹⁶ In contrast, lipid accumulation due to impaired lipolysis in adipose triglyceride lipase (ATGL)-deficient mice lead to a strong reduction of PPARα–regulated gene expression, defects in mitochondrial substrate oxidation, severe cardiac dysfunction and premature death. Treatment with a PPARα agonist reversed mitochondrial defects and normalized cardiac dysfunction, indicating that energy depletion due to impaired PPARα/PGC-1α signaling may have caused cardiomyopathy.^{97, 98} Collectively, these findings revealed a potential treatment for humans with neutral lipid storage disease and showed that endogenous or exogenous FA may first need to undergo esterification to TG and subsequent hydrolysis before becoming signaling lipids capable of activating PPARα. Of note, ATGL overexpression also protects from TG accumulation and cardiac dysfunction in models of obesity, diabetes, and pressure overload hypertrophy, by decreasing reliance of FAO on exogenous lipid and normalizing increased PPARα signaling.⁹⁹. Thus, the balance between oxidation of exogenous versus endogenous TG play an important role in the pathogenesis of lipotoxic cardiomyopathy.

Regulators of energy metabolism

Several transcriptional activators and coactivators participate in the regulation of cardiac energy metabolism. Transgenic mouse models of the PPARγ coactivators (PGC1s), PGC1α and PGC1β, provided insights into their role in cardiac physiology and metabolism. PGC-1 coactivators appear to be particularly important during postnatal growth since mice with combined deletion of PGC-1α and PGC-1β develop severe defects in mitochondrial dynamics and respiratory function, leading to contractile dysfunction, dilated cardiomyopathy and premature death .¹⁰⁰ During adulthood, although PGC1α may be required to maintain mitochondrial ATP-regenerating capacity by sustaining expression of genes involved in OXPHOS, TCA cycle and FAO, PGC1α KO mice do not develop contractile deficits under physiologic conditions.^{18, 21, 101} However, PGC-1α deficiency blunts the increase in ATP turnover and limits the increase in cardiac contractility following dobutamine stress and accelerates the progression to heart failure following TAC.^{18, 21} The latter phenotype may additionally be related to impaired induction of antioxidant gene expression in PGC1α-deficient mice as evidenced by attenuation of cardiac dysfunction in response to mitochondria-targeted ROS scavenging in this model. ¹⁰² Furthermore, PGC1β deficiency induces a shift towards decreased FAO and increased GO, which has been proposed to mediate the metabolic switch towards glucose utilization observed in pressure overload induced cardiac hypertrophy or models of heart failure.¹⁰¹ Mice deficient in PGC1β also show decreased OXPHOS expression and respiratory dysfunction, and accelerated progression towards heart failure in response to TAC, related to increased oxidative stress and the failure to increase glucose utilization.¹⁰³

Models with gain-of-function of PGC-1α provided evidence that PGC-1 coactivators can impressively stimulate mitochondrial biogenesis with a rapid and profound increase in mitochondrial mass, although accompanied by structural abnormalities of mitochondria and unexpected dilated cardiomyopathy.^{12, 104} This phenotype may be related to uncoordinated mitochondrial biogenesis due to strong activation of PGC-1α signaling or simply to excess mitochondrial mass within cardiomyocytes that spatially interferes with the contraction process. Thus, models with genetic manipulation of PGC coactivators support their role in regulating OXPHOS capacity, mitochondrial biogenesis, ROS homeostasis, and energy substrate preference in the heart. They also promote the postnatal increase in cardiomyocyte mitochondrial content and the switch from glycolysis towards predominant oxidation of FA for cardiac ATP regeneration. However therapeutic strategies utilizing PGC-1 stimulation to increase mitochondrial biogenesis in the heart may depend on the extent and speed of PGC-1 stimulation (reviewed in ¹⁰⁵).

Both PGC-1α and PGC-1β mediate effects on gene transcription and mitochondrial biogenesis by coactivation of transcription factors, predominantly including estrogen-related receptors (ERRα, ERRγ), nuclear respiratory factors (NRF1, NRF2) and mitochondrial transcription factor A (TFAm). While mice with homozygous deletion of TFAm show embryonic lethality due to severe mtDNA depletion and abolished OXPHOS activity, heterozygous deletion or conditional deletion in heart and muscle reduce mtDNA copy number, impaired transcription of mitochondria-encoded OXPHOS subunits and precipitated respiratory dysfunction, and dilated cardiomyopathy with lethality within one week of deletion. In contrast, mice with overexpression of human TFAm gene show increased mtDNA copy number, but no alterations of OXPHOS function or cardiac structure and function. Deficiency of additional factors regulating transcription and replication of mtDNA (mtDNA polymerase gamma, mitochondrial transcription termination factor 3, mitochondrial transcription factor 1 B) phenocopy TFAm deficiency such as embryonic lethality, mtDNA depletion, impaired OXPHOS expression, structural abnormalities, and cardiomyopathy, although not coactivated by PGC-1 coactivators (for specific details on these models see Supplemental Table S2).

NRF1 and NRF2 promote the expression of OXPHOS genes and frequently work in conjunction to optimize mitochondrial oxidative capacity. Mice lacking NRF1 established their essential role in the maintenance of mtDNA and respiratory chain function during early embryogenesis.¹⁰⁶ Mice with loss of NRF2 in mouse embryonic fibroblasts displayed decreased mitochondrial mass, ATP production, oxygen consumption, and mitochondrial protein synthesis without changing mitochondrial morphology or the expression of several genes previously reported to be NRF2 targets.¹⁰⁷ Effects of deletion of NRF1 or 2 in the heart and on cardiac energy metabolism have not been reported to our knowledge.

The orphan receptors, estrogen-related receptors (ERRs) transcriptionally regulate expression of genes involved in mitochondrial biogenesis and substrate utilization.¹⁰⁸ Mice with combined deficiency of ERRα and ERRγ show markedly decreased expression of FAO and OXPHOS genes, a strong defect in OXPHOS complex activities, display cardiac energy depletion, and die from dilated cardiomyopathy within the first month of life, illustrating that the ERR transcriptional pathway is essential for energy metabolism and maintenance of cardiac function.¹⁰⁹ In addition, expression of genes involved in FA oxidation, glucose uptake, TCA cycle, OXPHOS, and phosphate transfer is dysregulated in ERRα^−/− hearts at baseline and, in part, to a greater degree following pressure overload, thus identifying ERRα (a downstream component of PGC-1α signaling) as an important regulator of the bioenergetic and functional adaptation of the heart to hemodynamic stress imposed by pressure overload.¹⁰⁸ Mice with germline deletion of ERRγ display neonatal cardiac defects and die within 48 hours, which may be related to dysregulated expression of genes involved in OXPHOS and FA uptake and oxidation in the embryonic heart. These findings imply that ERRγ may govern the postnatal transition towards oxidative metabolism in the heart.¹¹⁰ Of note, cardiomyocyte-restricted ERRγ overexpression also induces early lethality due to dilated cardiomyopathy, which correlates with the extent of ERRγ overexpression. Expression of FA oxidation and OXPHOS genes was (unexpectedly) not significantly increased within the first three weeks of life in these mice but markedly repressed after onset of cardiac dysfunction. These studies underscore the importance of stoichiometry in transcription factor regulation of mitochondrial metabolism and biogenesis, as similarly observed in PGC-1 transgenic models.

PPARα is a nuclear hormone receptor and serves as the major transcriptional regulator of cardiac FA utilization by binding to response elements, predominantly in a complex with RXRα and coactivated by PGC-1α.¹¹¹ In global PPARα-KO mice, fasting induced TG accumulation in liver and heart due to inadequate induction of FA oxidation genes, thus defining a critical role for PPARα in a transcriptional regulatory response to fasting and identifying PPARα^−/− mice as a useful model of inborn and acquired abnormalities of human fatty acid utilization. In the heart, PPARα deficiency markedly impairs FAO and reciprocally increases glucose and lactate utilization, however this switch in substrate utilization is unable to maintain high-energy phosphate content under high workload conditions and thus leads to contractile failure, unless glucose uptake and utilization is further increased by overexpression of GLUT1 .¹¹² These findings from PPARα^−/− mice revealed that adult hearts with decreased capacity for FAO may be more susceptible to progressive deterioration of cardiac function during hemodynamic overload, and that metabolic interventions aiming at increasing glucose utilization may prevent energy depletion and cardiac deterioration in the setting of hemodynamic stress or pathological cardiac hypertrophy. Conversely, mice overexpressing PPARα specifically in cardiomyocytes showed increased myocardial FA uptake and oxidation, enhanced lipid content, a reciprocal decrease in glucose utilization, and impaired insulin-stimulated glucose uptake, thus serving as a model that mimics the cardiac metabolic phenotype commonly attributed to rodent or human diabetic cardiomyopathy.¹¹³

Other members of the PPAR family, PPARδ and PPARγ, also participate in myocardial lipid utilization. PPARδ deletion in cardiomyocytes impairs FA oxidation and increases myocardial lipid accumulation, leading to heart failure and reduced survival.¹¹⁴ This model revealed PPARδ as an important determinant of myocardial FA utilization and its necessity to maintain energy balance and cardiac function. In addition, mice with inducible PPARδ deletion in cardiomyocytes showed a strong suppression of mitochondrial biogenic signaling and evidence of oxidative stress, elucidating an additional role of PPARδ in mitochondrial protection and biogenesis in the heart.¹¹⁵ Germline knockout of PPARγ leads to embryonic lethality, however mice with cardiomyocyte-specific deletion of PPARγ are viable but develop cardiac hypertrophy with preserved systolic function, due in part to NFkB activation.¹¹⁶ While data on substrate metabolism are not available from this study, mice overexpressing PPARγ have increased myocardial TG uptake, FA oxidation, and lipid accumulation, but GLUT4 expression and glucose uptake were not decreased, and glycogen was even increased, accompanied by development of dilated cardiomyopathy.¹¹⁷ Thus, genetic modulation of PPARγ revealed that PPARγ is necessary to balance cardiac growth, and that overactivation results in cardiac glucolipotoxicity due to inappropriately high storage of TG and glycogen, potentially explaining cardiotoxic side effects of PPARγ agonist treatment in humans.

AMPK exists as a heterotrimeric complex consisting of a catalytic subunit α and two regulatory β and γ subunits and is activated by an increased AMP/ATP ratio or by phosphorylation of upstream regulators such as LKB1 or CaMKKIIb (reviewed in detail elsewhere¹¹⁸). In the heart, AMPK functions primarily as a metabolic sensor to coordinate anabolic and catabolic activities but can also affect other cellular processes such as mitochondrial function, posttranslational modifications, autophagy, ER stress, or apoptosis. AMPKα2-KO mice or mice with a kinase-dead or dominant-negative mutation of the α2 catalytic subunit show only mild or no defect in cardiac function under non-stressed conditions. Although AMPKα2 may be necessary to sustain mitochondrial cardiolipin content and OXPHOS complex I activity, effects on myocardial substrate utilization seem to be minor under non-stressed conditions, and unchanged ATP content and PCr/ATP ratio imply negligible compromise of myocardial energy status.^{119, 120} Instead, activation of AMPK in cardiomyocytes stimulates FAO, glucose utilization and mitochondrial biogenesis in cardiac muscle during conditions of unmatched energy demand such as exercise, ischemia, and pressure overload, indicating a predominant role of AMPK in regulating the adaptation of myocardial energetics during stress.^{120, 121} In contrast to the α2 subunit, transgenic mice with cardiomyocyte expression of a γ2 subunit (N488I) resulting in aberrant induction of AMPK activity, showed selective glycogen storage in concert with increased FAO and cellular glucose entry, and revealed that activation of mTOR and inactivation of FoxO signaling may underlie excessive hypertrophy in this model.^{122, 123} This model thus provided mechanistic insights into human glycogen storage cardiomyopathy induced by γ2 subunit mutation.

Glucose utilization:

Uptake and utilization of serum glucose requires trans-sarcolemmal transport of glucose via glucose transporters, as evidenced by data generated in mice with genetic manipulation of GLUT1 or GLUT4.^{124, 125} Increasing insulin-independent glucose uptake in mice with lifelong cardiomyocyte-selective overexpression of GLUT1 protects hearts with pressure-overload induced hypertrophy against contractile dysfunction and LV dilation, thereby providing experimental data supporting therapeutic strategies aiming at increasing glucose utilization in failing hearts.¹²⁶ Interestingly, short-term induction of GLUT1 at start of pressure overload was not sufficient to prevent LV contractile dysfunction, despite beneficial effects on mitochondrial, metabolic and structural remodeling, indicating that attenuation of structural and mitochondrial remodeling is offset by glucose-dependent mechanisms that limit contractility underscoring that optimal levels of increased glucose utilization may need to be defined to avoid detrimental effects of excess glucose availability.¹²⁷ In addition, increased FA availability as during high fat feeding may interfere with metabolic flexibility in hearts with increased GLUT1 expression and thus render the heart susceptible to contractile dysfunction.¹²⁸

Cardiomyocyte-specific GLUT4-KO abolishes insulin-stimulated glucose uptake but leads to compensatory increase in GLUT1-mediated basal glucose uptake.¹²⁹ While these metabolic changes only induce compensated hypertrophy, subjecting these mice to TAC exacerbated cardiac hypertrophy and promoted cardiac dysfunction, indicating that GLUT4 is required for maintenance of cardiac function and structure in response to chronic pressure overload.¹³⁰ Maintaining myocardial glucose and normalizing FA utilization via GLUT4 overexpression prevented cardiac dysfunction in transgenic db/db-hGLUT4 mice, providing evidence that altered energy metabolism contributes to cardiac dysfunction in models of type 2 diabetes.¹³¹ Combining cardiac GLUT4 overexpression with models of diabetes was also useful to reveal that increasing glucose utilization may suppress cardiac ketone utilization, and induce glucose- and O-GlcNAc-mediated mitochondrial dysfunction, thus identifying mitochondria as a target of glucotoxicity.^{53, 132}

Insights into the relationship between glycolytic flux and cardiac function were provided by mice with cardiomyocyte-specific overexpression of kinase-deficient PFK2, an important positive regulator of glycolysis. This mutation reduced glycolysis rates as expected but also revealed impaired insulin mediated Akt activation suggesting a direct link between glycolytic metabolism and Akt activation. Impaired insulin action could also be related to the accumulation of UDP-GlcNAc indicating increased flux into the hexosamine biosynthetic pathway which originates upstream of fructose-1,6-bisphosphate. The redirection of glycolysis intermediates in this model contributed to mild hypertrophy and impaired contractility.¹³³

The ability of PDKs (predominantly PDK4 in the heart) to inhibit glucose oxidation by PDH phosphorylation has been evidenced in mice with cardiomyocyte-specific overexpression of PDK4.¹³⁴ Conversely, mice with germline or conditional KO of PDK4 exhibit increased glucose oxidation and decreased FAO resulting in attenuation of both acute IR injury and postinfarction remodeling, with the latter beneficial effect being related to induction of cardiomyocyte proliferation. These studies underscore myocardial metabolism as a target for cardioprotective therapeutic strategies.

Oxidation of pyruvate by PDH requires mitochondrial pyruvate import. Recent studies revealed the significance of the mitochondrial pyruvate carrier (MPC) for TCA cycle metabolism of glucose-derived substrates and its role in cardiac structure and function. Cardiomyocyte-restricted deletion of subunit 1 of MPC (cmMPC1-KO) leads to age-related cardiac hypertrophy that progresses to dilated cardiomyopathy, associated with accumulation of upstream metabolites such as lactate, pyruvate and glycogen, and increased protein O-GlcNAc.¹³⁵ Interestingly, bypassing MPC1 by providing non-glucose substrates (KB, HFD) reversed the structural, metabolic and functional remodeling of cmMPC1-KO hearts by reducing flux of glucose-derived metabolites into pathways such as the hexosamine biosynthetic pathway that amplified ventricular remodeling. Similar findings were obtained in mice with cardiomyocyte-selective deletion of MPC2 (cmMPC2-KO) or mice with combined deletion or reduction of MPC1 and MPC2,^{136, 137} whereas overexpression of MPC attenuated isoproterenol-induced cardiac hypertrophy.¹³⁷ Moreover, increasing lactate conversion of pyruvate ameliorated ventricular remodeling in an isoproterenol heart failure model.¹³⁸ Collectively indicating that MPC-mediated mitochondrial pyruvate utilization is essential for the partitioning of glucose-derived cytosolic metabolic intermediates in the adaptation to myocardial stress.

Insulin and IGF signaling

Insulin signaling regulates glucose utilization in the heart and mediates its effects by binding to the insulin receptor and with less affinity to the IGF-1 receptor. Insulin receptor binding increases its autophosphorylation, which increases receptor tyrosine kinase activity for other substrates such as insulin receptor substrates (IRS) proteins. Consequently, IRS proteins and other binding partners activate a network of signaling components, including PI3K/AKT and MAP kinases.¹³⁹

Mice with cardiomyocyte-specific deletion of insulin receptors (CIRKO mice) enabled an assessment of myocardial insulin receptor signaling in the absence of confounding systemic metabolic derangements. Absence of cardiomyocyte insulin receptor signaling induced modest age-dependent cardiac dysfunction.¹⁴⁰ Underlying mechanisms may include cardiac metabolic insulin resistance, dysfunction and proteomic remodeling of mitochondria, and ROS-mediated mitochondrial uncoupling induced by FA exposure that can lead to impaired cardiac efficiency.³⁸ Based on these findings, it has been proposed that mitochondrial dysfunction and impaired cardiac efficiency observed in hearts of Type 2 diabetic rodents may be related to cardiac insulin resistance in these diabetes models.⁶¹ Insulin receptor and IGF-1R exhibit signaling redundancy in the heart.¹⁴¹ Indeed, mice with combined loss of insulin and IGF-1 receptors develop a lethal cardiomyopathy, indicating a partial protective effect of preserved IGF-1R signaling in CIRKO mouse hearts.

The significance of insulin signaling in regulating mitochondrial biology was confirmed in mice with cardiomyocyte-specific deletion of IRS1 and IRS2 that also develop mitochondrial dysfunction which precedes development of cardiac dysfunction. ¹⁴² Furthermore, this model provided valuable insight that insulin action in early life mediates the physiological postnatal suppression of autophagy, thereby linking nutrient sensing to postnatal cardiac development. Studies in single IRS knockouts revealed divergent roles in the regulation of cardiac size, however exercise-induced hypertrophy and the accompanying increase in oxidative capacity observed in wildtype mice were blunted both in mice with deletion of IRS1 or IRS2, indicating that both isoforms are required for the adaptation to chronic exercise training.

PI3K and AKT are downstream components of insulin signaling and regulate energy metabolism and cellular/organ growth. Mice with long-term gain-of-function and loss-of-function of PI3K revealed that PI3K increases cardiac growth.¹⁴³ Short-term activation of PI3K/AKT by insulin increases glucose uptake and decreases FAO, whereas sustained PI3K activation is sufficient to increase FA oxidative capacity independently of Akt and to suppress insulin-stimulated glucose uptake by impairing GLUT4 translocation.¹⁴⁴ These studies demonstrate that long-term metabolic effects of PI3K can be dissociated from downstream Akt signaling, and identified a role for PI3K in coordinating myocardial FAO capacity in exercise-induced cardiac hypertrophy. Data from mice with long-term activation or repression of Akt signaling demonstrated the potential of Akt to induce cardiac hypertrophy which may or may not progress to heart failure, partly related to the degree and duration of overexpression, insufficient induction of angiogenesis and repression of mitochondrial oxidative capacity.^{145, 146} Persistent Akt signaling may also be deleterious due to feedback inhibition of IRS and PI3K signaling.¹⁴⁷ Models with altered isoform-specific Akt expression informed Akt1 to promote physiological cardiac hypertrophy while antagonizing pathological hypertrophy, whereas Akt2 may not affect cardiac growth but instead mediate insulin-stimulated glucose utilization.¹⁴⁸

The IGF-1 receptor shares structural homology and overlapping downstream signaling components with the IR signaling cascade, however the affinity of the receptor is much higher for IGF-1 than for insulin. Cardiomyocyte-selective deletion of IGF-1 receptor prevents aging-associated cardiac hypertrophy due to inhibition of PI3K/AKT signaling, indicating a role for IGF-1 signaling in age-associated cardiac hypertrophy.¹⁴⁹ In addition, deletion of IGF-1 receptor prevents physiological cardiac hypertrophy in response to swimming exercise, whereas overexpression of IGF-1 receptor promotes physiological hypertrophy due to activation of the PI3K/AKT signaling pathway.¹⁵⁰ Detailed data on myocardial energy-metabolic rates are currently lacking. Of note, mice with combined deficiency of the insulin receptor and insulin-like growth factor 1 (IGF-1) receptor in cardiac and skeletal muscle develop early-onset dilated cardiomyopathy and die from heart failure within the first month of life, associated with a coordinated down-regulation of genes encoding for OXPHOS and FA oxidation proteins, suggesting that intact signaling via both receptors is required for normal cardiac metabolism and function

Ketone body and BCAA metabolism

Cardiac ketone metabolism has been investigated in mice with manipulation of BDH1 or SCOT expression. Deletion of BDH1 in cardiomyocytes impairs βOHB utilization and increases FAO leading to cardiac dysfunction, whereas BDH1 overexpression enhances both KB and FA utilization thereby attenuating cardiac hypertrophy and dysfunction induced by TAC.¹⁵¹ Cardiomyocyte-specific loss of SCOT increases FAO but does not affect cardiac structure or function during ketolytic challenge, whereas myocardial ROS levels and cardiac dysfunction were enhanced in response to TAC.¹⁵² Collectively, these findings highlight KB utilization as part of the coordinated adaptation to pressure overload and suggest KB utilization might protect against heart failure development, or may reflect induction of pathways that bypasses CPT1, such as induction of short chain fatty acid utilization, which also occurs in the failing heart.¹⁵³ Cardiomyocyte-specific deletion of BCAT revealed that BCAA accumulation due to defective BCAA oxidation induces cardiac hypertrophy, and that an accompanying decrease in BCKA levels may enhance myocardial metabolic insulin sensitivity.¹⁵⁴ Suppressing BCAA oxidation by deletion of mitochondrial PP2C (an activator of BCKDH) induced systolic dysfunction and exacerbated TAC-induced cardiac hypertrophy and dysfunction.¹⁵⁵ Collectively, BCAA oxidation seems to be required to sustain cardiac function and allow adaptation to pressure overload, implying that defective BCAA oxidation in failing hearts may causally contribute to heart failure development. Of note, another recent study demonstrated that the major fate of BCKA may not be oxidation but instead reamination to BCAA, thereby driving protein synthesis and cardiac hypertrophy via this mechanism. ¹⁵⁶ These findings allow the speculation that increased circulatory levels of BCAA and BCKA, as occurs in obesity, diabetes and early stages of heart failure, may contribute to hypertrophy development by diversion of BCKA to the reamination pathway.

Post-translational modifications

Sirtuins (SIRTs) are NAD⁺-dependent lysine deacetylases whose activity might modulate transcriptional pathways that regulate adaptations to changes in energy availability, aging, and stress resistance. While SIRT1, 2, 6 and 7 are predominantly localized in the nucleus and/or cytosol, SIRT3-5 are predominantly expressed in mitochondria. Mice expressing a mutant SIRT1 protein with a deleted catalytic domain or mice with germline SIRT1 deficiency, were severely growth retarded and developed cardiac developmental defects, which precluded post-natal survival.¹⁵⁷ In another study, germline knockout of SIRT1 led to dilated cardiomyopathy in adults, possibly due to morphological aberrations and functional defects of mitochondria. Interestingly, mild to moderate overexpression of SIRT1 retards aging-induced histological changes and LV dysfunction in the heart, whereas higher levels of SIRT1 expression leads to dilated cardiomyopathy, possibly through induction of mitochondrial dysfunction and ATP depletion.¹⁵⁸ Besides regulating mitochondrial oxidative metabolism, SIRT1 also participates in myocardial glucose utilization and insulin signaling via deacetylation and activation of Akt.

Several transgenic models have addressed the role of mitochondrial sirtuins. Global deletion of SIRT3 leads to progressive age-related deterioration of cardiac function in response to chronic pressure overload, which was proposed to result from increased acetylation and thus inhibition of various enzymes involved in energy substrate oxidation and OXPHOS, although the inhibitory effect of this acetylation has only been proven for few enzymes to date.¹⁵⁹ Conversely, overexpression of SIRT3 protects from angiotensin II-induced cardiac hypertrophy or doxorubicin-induced cardiomyopathy by suppressing cellular levels of ROS via FoxO3a-dependent expression of antioxidant proteins and by protecting from mitochondrial DNA damage, respectively, suggesting SIRT3 activation as a cardioprotective therapeutic strategy.¹⁶⁰ Opposite observations were made in animal models with altered expression of SIRT4 where overexpression of SIRT4 exacerbated and deletion of SIRT4 attenuated agonist-induced cardiac hypertrophy and dysfunction, possibly related to modulation of mitochondrial ROS via SOD2.¹⁶¹ Of note, SIRT4 also suppresses FAO, glucose oxidation and TCA cycle anaplerosis by deacetylation of MCD, delipoamidylation of PDH, and ADP ribosylation of glutamate dehydrogenase in extracardiac tissues, respectively (for review see ¹⁶²). However, detailed analysis of effects of SIRT4 on cardiac energy metabolism remain to be reported. Finally, systematic proteome profiling in mice lacking SIRT5 revealed numerous post translational modifications of mitochondrial proteins involved in mitochondrial metabolic pathways including desuccinylation and demalonylation of SIRT5 targets .¹⁶² While suppression of cardiac energetics has been proposed to underlie exaggerated hypertrophy and increased mortality in global SIRT5^−/− mice, the same group later reported that inducible and cardiomyocyte-restricted deletion of SIRT5 did not increase mortality or alter the hypertrophic and contractile response to TAC, leaving the role of SIRT5 in the heart to be further explored.¹⁶³

The role of SIRT6 and SIRT7 in cardiac metabolism have only been marginally explored. Cardiomyocyte-selective deletion of SIRT6 leads to cardiac hypertrophy and dysfunction, whereas overexpression protects from TAC-induced cardiac hypertrophy and dysfunction.¹⁶⁴ Mechanistically, SIRT6 may bind to and suppress the promoter of IGF signaling-related genes via interaction with c-Jun and deacetylation of histone 3, and may attenuate aging-associated alterations such as NAD⁺ depletion, mtDNA lesions or accumulation of 8-oxo-dG adducts.^{164, 165} In the context of MetS, SIRT6 overexpression preserves cardiac insulin sensitivity and attenuates cardiac hypertrophy, fibrosis and mitochondrial dysfunction, implying a potential for SIRT6 agonism in ameliorating diabetic cardiomyopathy. Mice with germline or cardiomyocyte-selective deletion of SIRT7 develop mild hypertrophy at older ages and exaggerated remodeling following TAC, likely related to p53-mediated excessive apoptosis and diminished resistance to oxidative/genotoxic stress.¹⁶⁶ Data on cardiac metabolism remain to be reported.

Another important posttranslational modification derived from glucose is the reversible attachment of O-linked N-acetylglucosamine moieties to Ser and Thr residues termed O-GlcNAc. Mice with deletion of O-GlcNAc transferase (OGT) and thus less O-GlcNAc demonstrated that this modification is necessary for cardiac maturation and induces cardiomyopathy.¹⁶⁷ Overexpression of OGT and increased O-GlcNAc also leads to dilated cardiomyopathy, likely due to disruption of OXPHOS complex I activity, indicating that importance of balanced O-GlcNAc to maintain cardiac integrity.¹⁶⁸ Of note, attenuation of O-GlcNAc by enhancing O-GlcNAcase activity is well tolerated and protects from pressure overload–induced pathologic remodeling, suggesting that this intervention may have therapeutic potential to attenuate cardiomyopathy. Of note, increasing flux through the hexosamine biosynthetic pathway by overexpression of the rate-limiting enzyme, glutamine:fructose-6-phosphate amidotransferase 1 (GFAT1) promotes cardiac hypertrophy through activation of mTOR, which appears to be mediated by increased O-GlcNAcylation.¹⁶⁹

Models of mitochondrial dynamics

Although mitochondria of the adult heart might be perceived as static organelles, models with genetic modulation of mitochondrial dynamics proteins advanced our understanding of fusion and fission for heart health. Studies in mice with deletion of both MFN1 and MFN2 exhibit impaired mitochondrial biogenesis, mtDNA replication, structural abnormalities, defects in mitochondrial gene transcription, and respiratory defects in the heart, overall leading to cardiomyopathy and severely reduced lifespan and indicating that mitochondrial fusion proteins are essential for mitochondrial remodeling during postnatal cardiac development and in the adult heart.^{170, 171} Comparative studies of mitochondrial fission- and fusion-defective hearts (deletion of MFN1, MFN2, or DRP1) suggested that mitochondrial fission may serve to segregate healthy and damaged components and to protect against transition pore-mediated cell death, whereas mitochondrial fusion may serve to repair and regenerate mitochondria, with both processes needed to be intact for proper elimination of impaired organelles and renewal of healthy mitochondria.¹⁷² Additional studies showed that MFN1/MFN2/DRP1 triple knockout mice have hearts with adynamic mitochondria that exhibit extended life span compared to their fission- or fusion-defective parents, but develop mitochondrial senescence and heart failure due to impaired mitophagy, mitochondrial superabundance, and sarcomere distortion.¹⁷³ Furthermore, proteolytic cleavage of the fusion protein, OPA1, drives mitochondrial fragmentation and cardiomyopathy development, accompanied by a switch from FA to glucose utilization, indicating an intimate link between mitochondrial morphology and energy metabolism in the heart.¹⁷⁴

Crosstalk between heart and peripheral metabolism

Although viewed as an obligate omnivore and a net consumer of metabolic energy, recent studies have shed novel insight into the regulation of systemic metabolism by factors that originate in the heart. One of the most dramatic descriptions of this concept was reported from the Olson laboratory, demonstrating that modulating the expression of MED13, a subunit of the mediator complex and its regulatory microRNA mir-208A, selectively in cardiomyocytes in vivo, led to significant regulation of systemic energy expenditure and modified susceptibility to diet-induced obesity and the metabolic syndrome,¹⁷⁵ via secreted factors, that remain to be characterized.¹⁷⁶ Knockout of the transcription factor KL5 in cardiomyocytes increased susceptibility to diet-induced obesity, which was prevented by simultaneous deletion in cardiomyocytes of the myokine fibroblast growth factor 21.¹⁷⁷ Inhibition of GRK2 activity in cardiomyocytes increased the susceptibility of mice to diet-induced obesity, whereas overexpression of GRK2 in cardiomyocytes, rendered animals resistant to diet-induced obesity. Metabolomics profiling revealed changes in myocardial and circulating levels of branch chain amino acids and endocannabinoid metabolites, which enhanced adipocyte differentiation in vitro.¹⁷⁸ Natriuretic peptides are secreted from hypertrophied and failing hearts. Adipose tissues express natriuretic peptide receptors,¹⁷⁹ activation of which may lead to browning of adipose tissue and increased energy expenditure.¹⁸⁰ Thus, natriuretic peptide secretion from the failing heart could contribute to altered systemic metabolism, and potentially cachexia in heart failure patients. Finally, novel mechanisms of crosstalk between peripheral metabolic organs and the heart, such as those mediated by exosomes have been recently described.^{181, 182}

Cardiac energy metabolism and models of HFrEF

Energy starvation contributes to the pathogenesis of HFrEF, linked in part to impaired ATP regeneration due to defects in mitochondrial oxidative capacity.¹⁸³ Underlying mechanisms of mitochondrial dysfunction include: increased mitochondrial ROS generation and oxidative stress which may induce mitochondrial DNA damage, damage to OXPHOS and other proteins, and lipid peroxidation, all of which may impair the ATP generating pathways.¹⁸⁴ Altered mitochondrial Ca²⁺ may further increase mitochondrial ROS production and alter mitochondrial NADH/NAD and NADPH/NADP ratios, thereby impairing the activity of Ca²⁺-dependent dehydrogenases (in particular of the TCA cycle) and triggering cell death pathways.¹⁸⁵ Excess mitochondrial fission leading to mitochondrial fragmentation, decreased antioxidative capacity, increased production of ROS, and cell death; ¹⁸⁶ sustained or impaired removal of damaged mitochondria by autophagy (i.e. mitophagy), thereby reducing the number of functional mitochondria or compromising the removal of damaged mitochondria, respectively, also contribute to mitochondrial dysfunction and its sequela including increased cell death in heart failure.¹⁸⁷

In addition, fuel preference is altered in the failing heart (Figure 4). Many studies demonstrated that during the development of pathologic hypertrophy or early in the course of cardiac failure, FA oxidation remains preserved, whereas in more severe stages of the disease, FA oxidation rates are mostly decreased.^{188, 189} These observations are consistent across various animal models of HF such as pressure overload-induced or MI-induced HF in rodents, pacing-induced HF in canines, or rats prone to develop hypertensive cardiomyopathy and HF. The decrease in FAO rates has been attributed to decreased expression of PPARα, RXRα, PGC1α, ERRs, and proteins of FA uptake and oxidation (e.g., CPT1b, MCAD, CD36, FATP). While these findings were also confirmed in human failing hearts, some studies reported no decrease or even an increase in FAO rates, likely related to differences in disease severity or to systemic metabolic comorbidities such as obesity and metabolic syndrome which secondarily influences cardiac substrate selection. ^{190, 191}

Figure 4: Simplified overview of energy metabolism in failing hearts.

Following transition from compensated hypertrophy to heart failure, both FA and glucose oxidation are impaired, mainly due to decreased transcriptional signaling (PPARα, PGC1α) and reduced expression of energy metabolic genes. While glucose oxidation and insulin stimulated glucose uptake is impaired in HF, glycolysis may be induced, driven in part by increased HIF1α signaling. A mismatch between glycolysis and glucose oxidation increases lactate levels and intracellular proton accumulation, and early intermediates of glycolysis drive cardiac remodeling by promoting flux into the PPP and HBP. KB oxidation is increased, whereas BCAA oxidation is decreased, resulting in myocardial and mitochondrial BCAA accumulation that promotes ROS generation and suppresses OXPHOS function. BCAA accumulation might also contribute to mTOR activation to increase cardiac hypertrophy. Downregulation and remodeling of OXPHOS subunits impair ATP synthesis and increases ROS generation, the latter promoting intramitochondrial oxidative damage, impaired Ca2+ handling, fibrosis, inflammation and cell death. ATP depletion may be further exaggerated by increased UCP expression and mitochondrial uncoupling. Collectively, these metabolic alterations contribute to cardiac remodeling and dysfunction in HF.

For detailed discussion of distinct alterations observed during compensated and decompensated stages of cardiac hypertrophy see refs^{2, 183, 190}.

Despite differences among studies and models, myocardial uptake and oxidation of glucose and thus the relative preference for ATP production is predominantly increased in the failing heart, in some models even starting at early stages of cardiac hypertrophy. While the adaptive or maladaptive nature of this shift towards preferential glucose utilization in the development of HF remains a subject of debate, it is generally presumed that switching to glucose may enable more oxygen-efficient production of ATP. However, at later stages of HF, glucose oxidation is generally decreased, due in part to decreased expression of PDH, MPC1, and pyruvate/alanine aminotransferases, or as a consequence of overall impairment of mitochondrial oxidative capacity and mitochondrial energetics. The impaired oxidation of all energy metabolic substrates leads to myocardial energy depletion. ¹³⁷ Furthermore, increased expression of UCP3 may induce mitochondrial uncoupling in some models (e.g., MI-induced HF), thereby compromising cardiac efficiency (=cardiac work/oxygen consumption). Decreased glucose oxidation and mitochondrial oxidative capacity is partially compensated by an induction of glycolysis which, only mildly increases ATP regeneration and may shift glucose into alternative pathways such as the pentose phosphate and hexosamine pathway, thereby further enhancing intracellular signaling that promotes adverse cardiac remodeling.¹⁹² In addition, uncoupling of glycolysis from glucose oxidation results in lactate production and proton accumulation in some HFrEF models, which renders glucose utilization less efficient and may secondarily increase cardiomyocyte Ca²⁺ accumulation. Increased glycolysis could be related in part to increased expression of GLUT1, and modification in the activity or expression of regulators of glycolytic flux, which may be mediated by increased HIF-1α levels as observed in rodent and human HF. ^{193, 194} HIF-1α is a hypoxia-inducible transcription factor, binding to the promoters of GLUT1 and other glycolytic pathway genes that contain HIF-1α response elements. A persistently activated hypoxic response in the heart as occurs during pressure overload or tachypacing may stabilize HIF-1α levels. HIF-1α transgenic overexpressing hearts develop age-dependent HF, accompanied by increased glucose uptake, increased expression of GLUT1 and PFK, a concomitant decrease of PPARα expression and a trend towards decreased FA oxidation.¹⁹⁴

HF is a ketosis-prone state. KB oxidation has been proposed to be more oxygen-efficient than oxidation of a mixture of FAs and glucose.¹⁹⁵ However, KB oxidation is a cataplerotic process (i.e. leading to loss of TCA cycle intermediates) and may ultimately compromise substrate oxidation and thus requires anaplerosis (oxaloacetate, malate) from oxidation of glucose or glucogenic amino acids to maintain TCA cycle activity. Plasma ketone levels and myocardial KB metabolism are increased in HF, but it remains a matter of debate whether enhanced myocardial KB utilization causes cardiac dysfunction, is caused by cardiac dysfunction, or simply is an epiphenomenon.⁵ The increased reliance on ketone body utilization in HF may result from increased systemic availability due to increased ketogenesis serving as an alternative fuel and possibly increasing oxygen-efficiency of ATP production.¹⁹⁶

Accumulation of amino acids has been observed in animal models of HF and in human failing hearts, which may result from impaired BCAA catabolism.¹⁹⁷ A transcriptome analysis of mice with pressure overload-induced HF revealed suppression of BCAA catabolic gene expression, in particular of pathways regulating catabolic pathways of valine, leucine, and isoleucine, including BCATm, BCKDH subunits E1α, E1β and E2, and the mitochondrially localized BCKDH phosphatase, protein phosphatase 2C (PP2C). As a result, BCKA metabolites accumulate in the myocardium, which can suppress respiration and induce superoxide production within mitochondria.¹⁵⁵ Importantly, mice lacking PP2C developed both accumulation of BCKA and systolic dysfunction and showed accelerated development of cardiac remodeling and dysfunction in response to pressure overload, suggesting that impaired BCAA catabolism contributes to HF pathogenesis. ^{155, 197}

Given the large number of studies performed in models of HFrEF, we provide an overview of frequently used HFrEF models and related characteristics of cardiac metabolism in Supplemental Table S3. For further details on these and other models of HFrEF, we refer to other excellent reviews.^{192, 198-200}

Novel models of HFpEF

HFpEF is a clinical syndrome with heterogenous etiologies, leading to distinct changes in myocardial structure and associated molecular pathways.²⁰¹ Comorbidities such as obesity, T2DM and hypertension are considered important causes of myocardial hypertrophy, ventricular stiffening, diastolic dysfunction and exertional dyspnea in human subjects suffering HFpEF. In human subjects and animal models of the MetS, the heart relies predominantly on FA utilization, coupled with impaired glucose utilization and mitochondrial oxidative defects, as described above. With regards to HFpEF pathophysiology, these models have however been criticized because they exhibit traits that are not necessarily observed in the human HFpEF syndrome. Importantly, some models may disqualify as bona fide HFpEF models since they may develop HFrEF at advanced stages of cardiac disease, which per definition is not compatible with a HFpEF diagnosis. For example, mice subjected to HFD, or aortic banding can ultimately lead to measurable systolic contractile defects, depending on the time point of investigation during follow-up and on variable factors (e.g., diet composition, degree of aortic constriction, genetic background, etc.), meaning that some models may only serve as a time-delimited model but not as bona fide model for human HFpEF. Further limitations of HFpEF modeling have been recently discussed in other excellent reviews.²⁰². These considerations need to be considered when evaluating animal models that are intended to mimic human HFpEF. Detailed phenotypes of models discussed below are summarized in Supplemental Table S4.

“Two hit” and “three hit” mouse models

Recently, the “two hit” model of HFpEF has been introduced, induced by the combination of HFD and inhibition of nitric oxide synthesis by L-NAME, which revealed increased body weight, visceral obesity, impaired glucose tolerance and insulin sensitivity, and increased blood pressure.²⁰³ As a result, hearts were reported to develop cardiac hypertrophy, diastolic dysfunction, myocardial fibrosis and increased pulmonary congestion, without evidence of systolic dysfunction up to 15 weeks of follow-up after treatment initiation, collectively matching a phenotype resembling that of clinical HFpEF. Metabolic studies revealed impaired glucose oxidation, unchanged expression of FAO genes, and impaired respiration using palmitoyl carnitine or pyruvate as substrates, indicating altered energy metabolism and mitochondrial dysfunction. As observed in models of HFrEF, treating the “two hit” mouse model with β-OHB protected against the progression of HF by increasing cardiac regulatory T cells by modulating NOX2/GSK-3β signaling, suggesting that this metabolic intervention may have merit in treating HFpEF by attenuating cardiac inflammation.²⁰⁴

In the “three hit” mouse model, an additional single bolus of desoxycorticosterone pivalate (DOCA) was administered in the last month of 13 months of HFD to accentuate hypertension and systemic inflammation.²⁰⁵ These mice develop myocardial fibrosis, inflammation, pulmonary edema, and diastolic dysfunction, in conjunction with increased body weight, reduced glucose tolerance, and elevated blood pressure. The study identified an interplay between mitochondrial protein hyperacetylation and NLRP3-related inflammation as a driver in the pathogenesis of HF. As in the “two hit” model, increasing circulating β-OHB also ameliorated the HF phenotype, likely by interrupting the vicious circuit of mitochondrial dysfunction and inflammation, reducing the acetyl-CoA pool (that drives protein acetylation) due to suppression of FA uptake and activation of citrate synthase. Of note though, while ejection fraction was not impaired, dP/dt was significantly reduced after 13 months, indicating subtle impairment in systolic function and indicating that this model may not be a bona fide model of HFpEF but instead shows a time-delimited HFpEF phenotype. Since mice from the ‘two hit” model were only observed over 15 weeks, development of systolic function can also not be excluded in this model.

Non-rodent models of HFpEF

Several non-rodent models of HFpEF have been recently reported. In a feline model, slow-progressive pressure overload induced by TAC using pre-shaped customized bands in 2 months-old kittens induced a progressive increase in concentric LV hypertrophy, enlargement and dysfunction of the left atrium, and diastolic dysfunction in the absence of systolic defects, subsequently leading to elevated LV filling pressures, elevated NT-proBNP levels, and pulmonary hypertension with impaired oxygenation and stiffening of the lungs.²⁰⁶ In addition, evidence for increased LV fibrosis, cardiomyocyte hypertrophy, perivascular cuff formation, impaired lung expansion, and alveolar-capillary membrane thickening was provided. Based on this cardiopulmonary analysis, this model might meet the diagnostic criteria of HFpEF. Studies of cardiac metabolism have not yet been performed.

Two models recapitulating the phenotype of HFpEF have been described in pigs where HFpEF was induced by inflation of aortic cuffs or by aortic banding with or without the addition of a western diet (high-fat/high-cholesterol).^{207, 208} Both models show increased body weight, plasma cholesterol, TG, insulin resistance, and aortic systolic BP, associated with cardiac hypertrophy and evidence of diastolic dysfunction, but no systolic defects. While the banding model exhibits mitochondrial dysfunction, mice with inflatable aortic cuffs showed increased PDH flux measured by magnetic resonance spectroscopy compatible with the typical switch from FA towards glucose utilization. As in rodent models, no data are currently available about systolic function after long-term treatment in these non-rodent models.

Translational potential of targeting cardiac energy metabolism

Preferential utilization of glucose

Promoting preferential use of glucose and non-FA substrates away from FAs, to increase oxygen efficiency of ATP regeneration is believed to mitigate or prevent energy depletion of the failing heart. Such metabolic modulation is associated with an improvement of the cardiac phosphocreatine:ATP ratio by up to 33%, which correlates with a parallel increase of left ventricular performance.²⁰⁹ Treatment with activators of glucose utilization (e.g. dichloroacetate) or inhibitors of FA utilization (e.g. trimetazidine, perhexiline, etomoxir, ranolazine) in animal studies, revealed improvements in ejection fraction and promoted structural reverse remodeling in experimental heart failure.²¹⁰ In humans, most evidence has been gained for trimetazidine, a drug with a favorable clinical safety profile causing partial inhibition of FA oxidation via long-chain 3-ketoacyl-CoA thiolase, increasing glucose oxidation, and inhibiting excessive glycolysis.²¹¹ While most clinical trials had limited patient size, meta-analyses of these trials revealed a reduction in hospitalization rates for cardiac causes and cardiovascular events, an improvement of ejection fraction, total exercise time, and NYHA function class, and a decrease in LV volumes and proBNP.^212-214 All-cause mortality was reduced in one analysis (RR 0.29) as well,²¹³ but remained non-significant despite clear trends (risk ratio < 0.50 in both analyses) towards a reduced risk ratio in the other two studies. Of note, an increase in cardiac function was associated with preferential cardiac glucose utilization in idiopathic dilated cardiomyopathy patients, confirming that trimetazidine indeed modulates cardiac metabolism in humans.²¹⁵ Of note, trimetazidine treatment not only improved systolic function and NYHA class, but also insulin sensitivity in HFrEF patients with concomitant insulin resistance or T2DM, suggesting that trimetazidine may also improve cardiac substrate utilization indirectly by improving systemic glucose metabolism. ²¹⁵ Despite some beneficial effects in small size clinical trials, less enthusiasm exists for other drugs with greater potency in shifting metabolism towards increased glucose utilization and away from FA metabolism. Oxfenicine an irreversible CPT1 inhibitor is not approved for human use. The potential risk of significant hepatotoxicity (etomoxir; reversible CPT1 inhibitor) or neuropathic toxicity (dichloroacetate; PDK inhibitor) has also precluded the use of these agents.^{216, 217} Increasing the coupling of glycolysis to glucose oxidation, thereby preventing accumulation of glycolytic intermediates that could promote LV remodeling and also reducing proton accumulation thereby increasing the efficiency of ATP production from glucose, may be another promising concept which could be achieved by inhibition of either MCD or ACC. Clinical data on these agents are not yet available.¹⁹⁸

Modulation of BCAA and KB utilization

Accumulation of amino acids has been observed in HF, which may result from impaired BCAA catabolism.^{197, 218} Enhancing BCAA catabolism by pharmacological inhibition of BCKDH kinase blunted systolic and diastolic dysfunction after pressure overload,²¹⁹ suggesting that enhancement of BCAA catabolism (by activation of catabolic enzymatic activity) could serve as a potentially efficacious treatment option in HF.^{197, 219} Cardiac and serum levels of BCAA are also increased in obesity, IR and T2DM, and circulating levels of BCAAs are independently associated with incident HF in subjects with T2DM.²²⁰ Dietary restriction of BCAAs has been shown to shift cardiac metabolism towards FA catabolism relative to glucose metabolism and eliminated a 2-fold increase in myocardial triglycerides. This also improved insulin resistance both in obese and/or diabetic rodents and humans, suggesting that both direct and indirect mechanisms (secondary to improved systemic metabolism) may modulate cardiac energy metabolism.^{221, 222} While increasing utilization or dietary restriction of BCAAs in HF and T2DM may exert beneficial effects, more studies in humans are required to further evaluate this therapeutic strategy.

Several studies evaluated increasing KB oxidation as a therapeutic strategy in the failing heart. A recent study demonstrated that a ketogenic diet (high fat/low carbohydrate) attenuated the extent of adverse cardiac structural remodeling in a combined TAC/MI model, and that chronically increasing delivery of ketones (4 weeks) in the canine tachypacing model of HF markedly improved cardiac function and remodeling.²²³ In this canine model of HF, ketone body uptake was increased compared to controls, and was further doubled under increased KB delivery. The beneficial response is possibly due to increased mitochondrial respiratory efficiency when utilizing ketones as a substrate. Also, direct application of βOHB (as ketone salt) by intravenous infusion increased cardiac output in HFrEF patients, and a dose-response relationship between β-OHB levels and cardiac output was shown.²²⁴ Furthermore, genetically induced elevation of circulating ketone bodies, without a coinciding reduction in glucose levels, protected against the development of pressure-overload induced HF by impairing NLRP3 inflammasome activation, suggesting that ketone bodies may prevent cardiac inflammation.²²⁵ While these studies may suggest that increased ketone utilization in HF may be an adaptive stress response, increasing myocardial KB utilization did not improve cardiac function in all studies.²²⁶ The durability of increasing ketone utilization to improve cardiac dysfunction in heart failure remains to be determined. Moreover, the method of delivery (duration, route of application, ketone chemistry) to increase myocardial ketone oxidation, and the underlying mechanisms how ketones may benefit the failing heart remain to be elucidated. Of note, certain routes of ketone administration (e.g., infusions) might not be feasible for clinical practice.

Weight loss strategies and impact on cardiac metabolism in heart failure

Weight loss, achieved by caloric restriction (CR) or intermittent fasting, improves systemic metabolism to mitigate cardiovascular risk factors that could reduce the onset and progression of CVD. CR or time-restricted feeding may induce sustained weight loss, improve insulin sensitivity, lower blood pressure, improve CRP levels, and lower oxidative stress in non-obese, obese, prediabetic or diabetic subjects.^{227, 228} CR has been shown to increase peak MVO₂ in obese patients with HFpEF, and to improve diastolic dysfunction and to lower myocardial TG content in obese and diabetic subjects.^{229, 230} In rodents, CR and intermittent fasting also improved LV remodeling in HFrEF models.²³¹ Studies in mice with pressure-overload hypertrophy and subjected to HFD suggest that these beneficial cardiac effects correlate with altered cardiac energy metabolism, including enhanced cardiac insulin signaling and insulin-stimulated glucose oxidation, decreased cardiac FA oxidation, reduced AMPK signaling and acetylation of ß-oxidation enzymes.²³² In diabetic Long-Evans Tokushima Otsuka rats or in db/db mice, beneficial effects of CR may be related to increased SIRT1 expression, AMPK phosphorylation, FoxO1 phosphorylation, and PGC1α signaling, that correlate with improved diastolic and systolic LV function, and reverse remodeling.²³³ While weight loss strategies seem to indirectly improve cardiometabolic health, reduced compliance for lifelong CR and the potential risk of inducing essential nutrient deficiency, are limitations of weight loss-based approaches to heart failure , which would require additional study. Intermittent fasting could potentially be more feasible but remains to be studied in the context of heart failure.

Resveratrol

Resveratrol is a natural polyphenol that mimics effects of dietary restriction and/or every-other-day feeding. In rodent models of HF, resveratrol treatment increased median survival, attenuated cardiac fibrosis and structural remodeling, and improved or prevented cardiac dysfunction.²³⁴ While resveratrol attenuated oxidative stress, the functional improvements may be related to restoration of glucose oxidation (by shifting substrate preference towards glucose), restoration of OXPHOS subunit expression and mitochondrial respiratory function, normalization of AMPK phosphorylation, increased mitochondrial biogenesis and enhanced PPARα and SIRT1 signaling.^{235, 236} Specific underlying mechanisms remain incompletely understood. In rodent models of the metabolic syndrome, administration of Resveratrol protects against the development of diet-induced insulin resistance and obesity, and lowers circulating concentrations of triglycerides, FAs, insulin and leptin. These effects on systemic metabolism would be expected to beneficially affect cardiac energetics, structure and function, as observed in rat models of diabetic cardiomyopathy subjected to resveratrol treatment.²³⁷ While administration of resveratrol in humans is clinically feasible and safe, effects on diabetic or failing hearts in human subjects have not yet been reported.

SGLT2 inhibitors (SGLT2i)

SGLT2i, which are approved for the treatment of diabetes, lower blood glucose levels by inhibiting renal glucose reabsorption via inhibition of the sodium glucose cotransporter. Importantly, SGLT2i treatment decreased HF hospitalization and/or cardiovascular death in patients with HFrEF with and without T2DM.²³⁸ Inhibition of SGLT2 results in a mild hyperketonemia, which led to the proposal that improved myocardial ATP regeneration due to increased KB oxidation may be responsible for beneficial effects of SGLT2i. However, while empagliflozin treatment improves cardiac respiratory function and ATP production, and attenuated cardiac dysfunction and remodeling in rodent models of T2DM, KB oxidation remained unaffected by the treatment.^{239, 240} In MI-induced HF in pigs, empagliflozin treatment resulted in a switch towards utilization of KB, FA, and BCAA, which was associated with increased myocardial ATP content, amelioration of cardiac function and structure, and improved cardiac efficiency,²⁴¹ however this was not observed in pressure-overload induced HF in mice, despite improved cardiac function.²⁴² Thus additional research is required to understand if and how SGLT2i may exert beneficial cardiovascular effects by modulation of cardiac energy metabolism.

Conclusions

This comprehensive review and its supplementary materials, provides a comprehensive overview of a large body of work examining cardiac metabolism in a range of animal models ranging from those with altered systemic metabolism to those with specific perturbations in metabolic pathways or their signaling regulators. We have also summarized the cardiac adaptations to metabolic conditions such as diabetes and obesity, the metabolic and mitochondrial maladaptations that characterize heart failure and the interaction between these conditions. Lessons learned from these studies include the multiple and redundant regulatory mechanisms that exist in the heart to ensure metabolic homeostasis. The importance of maintaining metabolic flexibility has been underscored. Constraints on these mechanisms may contribute to functional maladaptations that lead to the progression of heart failure. The models reveal that in addition to ATP generation, metabolic substrates also regulate multiple signaling pathways that regulate myocardial homeostasis, disturbances of which may contribute to myocardial dysfunction. To the extent that animal models have provided significant and novel insights into metabolic regulation in health and disease, future studies should focus on determining the relevance of these findings to humans and evaluate how these findings can be translated into therapies for cardiovascular disease and heart failure.

Abbreviations:

4EBP1

4E-binding protein

8-OHdG

8-hydroxydeoxyguanosine

AA

amino acid

ABCA

ATP-binding cassette transporter

Ac

acetylated

ACAA2

acetyl-CoA acyltransferase 2

AcAc

acetoacetate

ACAD

Acyl-CoA dehydrogenase

ACADL

long-chain ACAD

ACADM

medium-chain ACAD

ACAT

acetyl-CoA acetyltransferase

ACC

acetyl-CoA carboxylase

ACO

ACC oxidase

ACON

aconitate hydratase

ACOT

acyl-CoA thioesterase

ACS

acyl-CoA synthetase

AcylCN

acylcarnitine

ADP

adenosine diphosphate

ADP/O

ADP:oxygen ratio

ADRP

adipose differentiated related protein

AGPAT

1-acylglycerol-3-phosphate O-acyltransferase

AKAP

A-kinase anchoring protein

aKGDH

α-ketoglutarate dehydrogenase

AKT

protein kinase B

ALT

alanine aminotransferase

AMP

adenosine monophosphate

AMPK

adenosine monophosphate-activated protein kinase

ANP

atrial natriuretic peptide

ANT

adenine nucleotide translocase

AOX

alternative oxidase

ApoB

apolipoprotein B

ASCL

acyl-CoA synthetase, long chain

ASNS

asparagine synthetase

ATGL

adipose triglyceride lipase

ATP

adenosine triphosphate

ATP/O

ATP:oxygen ratio

ATP5A1

ATP synthase α-subunit

ATP5F1

ATP synthase F1

ATP5PD

ATP synthase peripheral stalk subunit D

ATPsyn

ATP synthase

BCAA

branched chain amino acid

BCAT

branched-chain aminotransferase

BCATm

branched-chain aminotransferase, mitochondrial

BCKDH

branched chain alpha-keto acid dehydrogenase

BDH

β-hydroxybutyrate dehydrogenase

BPHx

hypertensive BPH/2J Schlage

BT2

3,6-Dichlorobenzo[b]thiophene-2-carboxylic acid

BW

body weight

CASP

caspase

CD36

fatty acid transporter

CGP

choline glycerophospholipid

ChE

cholesteryl esters

CHO

carbohydrate

CK

creatine kinase

CL

cardiolipin

CM

cardiomyocyte

CoA

coenzyme A

complex I

NADH-Q oxidoreductase

complex II

succinate dehydrogenase

complex III

Coenzyme Q:cytochrome c reductase

complex IV

cytochrome c oxidase

complex V

F0F1 ATP synthase

COX

cytochrome c oxidase

CPT

carnitine palmitoyltransferase

CrAT

carnitine polmitoyl transferase

CS

citrate synthase

CT

CTP:phosphocholine cytidylyltransferase

CYT

cytochrome

DAG

diacylglycerol

DBP

diastolic blood pressure

DCA

dichloroacetate

DGAT

diacylglycerol acyltransferases

DHQ

dihydroquercetin

DOCA

deoxycorticosterone acetate

DRP

dynamin-related protein

DSS

Dahl salt-sensitive

EC

endothelial cell

ECH

enoyl-CoA hydratase

ECHA

trifunctional enzyme α subunit

eEF

eukaryotic elongation factor

EF

embryonic fibroblast

EKG

electrocardiogram

EP

ethanolamine phospholipids

ER

endoplasmic reticulum

ERK

extracellular Receptor Kinase

ERR

estrogen-related receptor

ESC

embryonic stem cell

ETC

electron transport chain

ETFA

electron transfer flavoprotein alpha

F6P

fructose 6-phosphate

FA

fatty acid

FABP

FA binding protein

FABPpm

FABP, plasma membrane

FADH2

flavin adenine dinucleotide

FAO

FA oxidation

FAS

FA synthase

FATP

FA transport protein

FBP

fructose-1,6-bisphosphatase

FFA

free FA

FGF

fibroblast growth factor

FIS

fission

FOXO

forkhead box O

Fru-2,6-P(2)

fructose 2,6-bisphosphate

G1P

glucose-1-phosphate

G6P

glucose-6-phosphate

GABA

gamma-aminobutyric acid

GABPA

GA repeat-binding protein α

GATP

ATP free energy

GCN5L1

GCN5-like protein

GDF

growth differentiation factor

GLDH

glutamate dehydrogenase

GLUT

glucose transporter

GPAM

glycerol-3-phosphate acyltransferase, mitochondrial

GPAT

glycerol-3-phosphate acyltransferase

GS

glycogen synthase

GSH

glutathione

GSK

glycogen synthase kinase

H202

hydrogen peroxide

HADH

3-hydroxyacyl-CoA dehydrogenase

HbA1c

hemoglobin A1c

HFD

high-fat diet

HFHS

high-fat/high-sugar

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

HIF

hypoxia-inducible Factor

HK

hexokinase

HMG-CoA

β-Hydroxy β-methylglutaryl-CoA

HMGB

high mobility group box protein

HMGSC

3-hydroxy-3-methylglutaryl-CoA synthase

HOMA-IR

homeostasis model assessment-estimated insulin resistance

HR

heart rate

HR-LPL

heparin-releasable LPL

HSL

hormone-sensitive lipase

HSP

heart shock protein

HW

heart weight

ICHD

isocitrate dehydrogenase

IFM

interfibrillar mitochondria

IGF

insulin-like growth factor

IGF1R

IGF-1 receptor

IL

interleukin

IL1RL

interleukin 1 receptor-like

IR

insulin receptor

IRS

IR substrate

JNK

c-Jun N-terminal kinase

KB

ketone body

KEAP

kelch-like ECH-associated protein

KI

knock-in

KIV

ketoisovalerate

KLF15

Krüppel-like factor 15

KO

knockout

L-CPT

CPT liver isoform

L-NAME

Nitro-l-arginine methyl ester hydrochloride

LAD

left anterior descending

LC

long-chain

LC-MS/MS

liquid chromatography–mass spectrometry

LCAC

long-chain acylcarnitine

LCAD

long-chain acyl CoA dehydrogenase

LCFA

long-chain FA

LCTH

long-chain 3-ketoacyl-CoA thiolase

LDH

lactate dehydrogenase

LDL

low-density lipoprotein

LDLR

low-density lipoprotein receptor

LIVCS

BCAA:cation symporter family

LpL

lipoprotein lipase

LPS

lipopolysaccharide

LV

left ventricle

LVRI

left ventricular remodelling index

lycoPC

lysophosphatidylcholines

LysAc

lysine acetylation

MAP

mean arterial pressure

MAPK

mitogen-activated protein kinase

MCAD

medium-chain acyl-CoA dehydrogenase

MCD

malonyl-CoA decarboxylase

MCT

monocarboxylate transporter

MEF

mouse embryonic fibroblasts

MetS

metabolic syndrome

MFN

mitofusin

MHC

myosin heavy chain

MI

myocardial infarction

MMP

mitochondrial membrane potential

MnSOD

manganese superoxide dismutase

mPTP

mitochondrial permeability transition pore

MRS

magnetic resonance spectroscopy

mtDNA

mitochondrial DNA

MTE

mitochondrial thioesterase

MTERF

mitochondrial transcription termination factor

MTHFD

methylenetetrahydrofolate dehydrogenase

mTOR

mammalian target of rapamycin

MUFA

monounsaturated fatty acids

MVo2

myocardial O2 consumption

NAD

nicotinamide adenine dinucleotide

NADH

reduced NAD

NADPH

nicotinamide adenine dinucleotide phosphate

NDUFS

NADH-ubiquinone oxidoreductase, mitochondrial

NDUFS3

mitochondrial complex I subunit

NDUFV

NADH dehydrogenase [ubiquinone] flavoprotein, mitochondrial

NMNAT

nicotinamide mononucleotide adenylyltransferase

NRF

nuclear respiratory factor

NYHA

New York Heart Association

O-GlcNA

O-GlcNAcylation

OE

overexpression

OGA

O-GlcNAcase

OGG

oxoguanine-DNA glycosylase

OGT

O-GlcNAc Transferase

OPA

optic atrophy

OXPHOS

oxidative phosphorylation

PC

phosphatidylcholines

PCr

phosphocreatine

PDK

pyruvate dehydrogenase kinase

PDL1

programmed death-ligand

PET

positron emission tomography

PFK

phosphofructokinase

PFKFB

6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3

PG

phosphatidylglycerol

PGC

PPARy co-activator 1a

PGI

phosphoglucose isomerase

PGK

phosphoglycerate kinase

Pi

phosphatidylinositol

PI3K

phosphoinositide 3-kinase

PKC

protein kinase C

PL

phospholipid

PLA

lipoprotein-associated phospholipase A

PLB

phospholamban

PLIN

perilipin protein

POLGA

DNA polymerase gamma, subunit A

PP

pulse pressure

PP2A

protein phosphatase 2A

PP2Cm

protein phosphatase 2C , mitochondrial

PPAR

peroxisome proliferator-activated receptor

PRAS

proline-rich Akt substrate

PRC

PGC-1-related coactivator

proBNP

pro b-type natriuretic peptide

PRX

peroxiredoxin

PTP

permeability transition pore

PTX3

pentraxin-3

RAAS

renin-angiotensin-aldosterone system

RNA

ribonucleic acid

ROS

reactive oxygen species

RV

right ventricle

RXR

retinoid X receptor

RyR

ryanodine receptor

S6

ribosomal protein s6

S6K

S6 kinase

S6K1

ribosomal protein S6 kinase beta

SBP

systolic blood pressure

SCAD

short-chain

SCOT

succinyl-CoA:3-oxoacid-CoA transferase

SERCA

sarcoplasmic reticulum Calcium-ATPase pump

SFA

saturated fatty acid

SGLT2i

sodium-glucose cotransporter-2 inhibitor

SH

spontaneously-hypertensive

SHR

spontaneously hypertensive rat

SIRT

sirtuin

SLC

solute carrier family

SM

sphingomyelins

SP

specificity protein

SREBP

sterol element binding regulatory proteins

SSM

subsarcolemmal mitochondria

STAT

signal transducer and activator of transcription

STZ

streptozotocin

T2DM

type 2 diabetic

TAC

transverse aortic constriction

TCA

tricarboxylic acid cycle

TFAm

mitochondrial transcription factor A

TFB

transcription factor B

TFB1M

TFB1M transcription factor B1

TG

triglyceride

TGF

transforming growth factor

TIGAR

TP53-induced glycolysis and apoptosis regulator

TnC

troponin C

TNF

tumor necrosis factor

TnT

troponin T

TRIB

tribbles homolog

TRPC

transient receptor potential cation channel

TTN

titin

UCP

uncoupled protein

UDP

uridine diphosphate

UDPG-PPL

UDP-glucose pyrophosphorylase

UNSAT

unsaturated fat

UPR

unfolded protein response

UQCRC

ubiquinol-cytochrome c reductase core protein

VDAC

voltage-dependent anion-selective channel

VEGF

vascular endothelial growth factor

VLCAD

very long-chain acyl-CoA dehydrogenase

VLDL

very low density lipoprotein

WAT

white adipose tissue

WD

western diet

YME1L

ATP-dependent metalloprotease

ZDF

Zucker diabetic fatty

ZF

Zucker fatty

α-KGDH

alpha-ketoglutarate dehydrogenase

βOHB

β-hydroxybutyrate