|Home | About | Journals | Submit | Contact Us | Français|
Decreased fatty acid oxidation (FAO) with increased reliance on glucose are hallmarks of metabolic remodeling that occurs in pathological cardiac hypertrophy and is associated with decreased myocardial energetics and impaired cardiac function. To date, it has not been tested whether prevention of the metabolic switch that occurs during the development of cardiac hypertrophy has unequivocal benefits on cardiac function and energetics.
Since malonyl CoA production via acetyl CoA carboxylase 2 (ACC2) inhibits mitochondrial fatty acid transport, we hypothesized that mice with a cardiac-specific deletion of ACC2 (ACC2H−/−) would maintain cardiac fatty acid oxidation (FAO) and improve function and energetics during the development of pressure-overload hypertrophy.
ACC2 deletion led to a significant reduction in cardiac malonyl CoA levels. In isolated perfused heart experiments, left ventricular (LV) function and oxygen consumption were similiar in ACC2H−/− mice despite an ~60% increase in FAO compared to controls (CON). After 8 weeks of pressure-overload via transverse aortic constriction (TAC), ACC2H−/− mice exhibited a substrate utilization profile similar to sham animals while CON-TAC hearts had decreased FAO with increased glycolysis and anaplerosis. Myocardial energetics, assessed by 31P NMR spectroscopy, and cardiac function were maintained in ACC2H−/− after 8 weeks of TAC. Furthermore, ACC2H−/−-TAC demonstrated an attenuation of cardiac hypertrophy with a significant reduction in fibrosis relative to CON-TAC.
These data suggest that reversion to the fetal metabolic profile in chronic pathological hypertrophy is associated with impaired myocardial function and energetics and maintenance of the inherent cardiac metabolic profile and mitochondrial oxidative capacity is a viable therapeutic strategy.
Fatty acid oxidation (FAO) is a major energy source for the adult mammalian heart. Decreased FAO contributes to the reappearance of the fetal metabolic pattern in hypertrophied and failing hearts that leads to increased reliance on glycolysis combined with upregulation of anaplerosis to maintain tricarboxcylic acid cycle (TCA) flux.1–5 Although a shift towards carbohydrate metabolism slightly improves myocardial oxygen efficiency, this metabolic profile is inefficient in utilizing carbon substrates for ATP production during increased energy demand, leading to impaired myocardial energetics and depletion of contractile reserve.6,7 Restoration of FAO with PPARα agonists in hypertrophied and failing hearts has been unfruitful,8,9 likely due to broad effects of these compounds on lipid uptake and metabolism. However, recent reports showed the benefit of high fat diets in certain models of heart failure.10,11 Since increased myocardial FAO has been implicated in the development of metabolic cardiomyopathy in obesity and diabetes,12,13 it remains controversial whether enhancing FAO during the development of pathological cardiac hypertrophy will prevent metabolic remodeling and preserve myocardial energetics and function.
The rate-limiting step of FAO is the import of long chain fatty acids (FA) across the mitochondrial membrane through carnitine palmitoyl transferase I (CPT1). This action is strongly inhibited by malonyl CoA, which is formed by the carboxylation of acetyl CoA via acetyl CoA carboxylase (ACC).14 Deletion of ACC2, the primary ACC isoform in oxidative tissues, led to increases in whole-body and tissue specific FAO.15–17 Although originally shown to improve insulin sensitivity and resist against diet-induced obesity and diabetes,18,19 follow-up studies from two independent groups found that deletion of ACC2 upregulated mitochondrial FAO without altering overall energy homeostasis,17,20 Nonetheless, these observations are consistent with past reports demonstrating that malonyl CoA produced by ACC2, is a target for modulating substrate preference of oxidative tissues.21,22
A prior report showed hearts of ACC2-null mice displayed increased oxidation of both glucose and FA with reduced heart size but unaltered cardiac function.16 This strain of ACC2-null was leaner with altered whole body energy homeostasis.15,19 It is unclear whether the cardiac phenotype in those studies is the result of altered systemic metabolism or due to elevated cardiac FAO. We generated a mouse model with a cardiac-specific deletion of ACC2 in order to interrogate the specific effects of reduced cardiac malonyl CoA levels on cardiac substrate metabolism in normal and hypertrophied hearts. Our results show that deletion of ACC2 results in a shift of substrate oxidation to FAO without negatively impacting cardiac function in the long term. Moreover, ACC2 deletion prevents metabolic reprogramming and sustains myocardial energetics and function in pathological cardiac hypertrophy.
An expanded Methods section is provided in the Online Data Supplement.
Animal studies were approved by the Harvard Medical Area Standing Committee on Animals or University of Washington Institutional Animal Care and Use Committee. ACC2 flox/flox (ACC2f/f) mice on a C57/129 background20 were mated with αMHC-Cre mice to yield four genotypes: ACC2f/WT, ACC2f/f, ACC2f/WT-Cre+ (ACC2−/+), and ACC2−f/f−Cre+ (ACC2H−/−). ACC2H−/− were mated with ACC2f/f to produce both study and control littermates. Mice were kept on a 12-hour light/dark cycle with water and food ad libitum.
Detection and quantification of malonyl CoA levels was adapted from a previously described methods using LC-MS/MS.28,29 For acylcarnitines species, frozen ventricular tissue from non-perfused hearts were powdered and homogenized in deionized water.30 Acylcarnitines were quantified using commercially available labeled standards and normalized to protein concentration. For global metabolite profiling, ventricular tissue was extracted in 2:1 chloroform:methanol and analyzed by GC × GC-TOFMS.31 The previously developed F-ratio method was used to analyze all samples at each mass channel. Detection of additional metabolic features was made via Peak Table analysis (LECO ChromaTOF software version 3.2, LECO Corp., St. Joseph, MI, USA). Peak areas of target analytes were measured by PARAFAC analysis.32 (See Online Supplement for additional details).
All data are presented as mean ± SEM. Student’s t-test was used for two group comparisons. One-way ANOVA was performed for multiple-group comparisons at single-time point. Two-way ANOVA was used for multiple group comparisons with 2 factors (i.e., substrate utilization). ANOVA with repeated measures was used for multiple group comparisons over multiple time points. Bonferroni post hoc analysis was used for all ANOVAs. Analyses were performed with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). All results were tested at the P < 0.05 level of significance.
ACC2 flox/flox (ACC2f/f) mice were generated as previously described,20 and cardiac-specific deletion of ACC2 (ACC2H−/−) was achieved by cross-breeding ACC2f/f with αMHC-Cre mice. No differences were found in serum levels of glucose, insulin, free FA, and triglycerides across all genotypes (Online Table I). Furthermore, ACC2H−/− had no significant differences in heart weight (HW) or HW relative to body weight (BW) or tibial length (TL) (Online Table II), suggesting a possible systemic and/or strain effect in the previously observed smaller heart size in ACC2-null mice.16,20 Western blotting confirmed the cardiac-specific deletion of ACC2 (Figure 1A, 1B). Since physical characteristics and cardiac protein levels of ACC2 in both ACC2f/WT and ACC2f/f were similar to wild type C57BL6 mice, ACC2f/f mice were used as controls (CON). Because the ACC1 isoform is present in the heart at low levels (~10%) and is difficult to detect via Western blotting,33 we measured changes in gene expression to confirm a negligible effect of ACC2 deletion on ACC1 (Figure 1C). Deletion of ACC2 resulted in a significant decrease (~50%) in cardiac malonyl CoA levels (Figure 1D).
In isolated Langendorff-perfused heart experiments, ACC2H−/− demonstrated an ~60% increase in FAO as compared to CON at 8–10 weeks of age (P< 0.05); Figure 2A), that was concomitant with a reciprocal reduction in the contribution of glucose oxidation ( P<0.05); Figure 2A). Despite enhanced FAO, cardiac triglycerides and glycogen levels remained unchanged in ACC2H−/− (Figure 2B, 2C). A recent report suggested that increased FAO is associated with mitochondrial overload of acylcarnitines and decreased TCA cycle intermediates that contribute to the development of insulin resistance in skeletal muscle of mice fed a high fat diet.30 Thus, we assessed acylcarnitine species in non-perfused heart extracts from CON and ACC2H−/−. No significant differences were noted in the concentrations of short-chain (C3, C4, C5), medium chain (C8), or long-chain (C14, C16) acylcarnitine species (Figure 2D). In addition, expression of genes involved in glucose metabolism, FA metabolism, and mitochondrial biogenesis were not significantly different in ACC2H−/− as compared to CON (Figure 2E). Overall these results suggest that cardiac-specific deletion of ACC2 increases cardiac FAO without overwhelming changes of substrate metabolism.
Since the 60% increase of FAO in ACC2H−/− renders a substrate oxidation profile similar to a diabetic heart,12 we determined whether chronic elevation of mitochondrial FAO led to detrimental consequences during normal aging. First, we verified that 12 month old ACC2H−/− mice maintained elevated FAO and had a similar substrate utilization profile as 2 months of age (Figure 3A). Next we performed serial echocardiography to evaluate cardiac function and morphology in ACC2H−/− and CON up to 12 months of age. Fractional shortening and heart rate were similar in CON and ACC2H−/− at 2, 6 and 12 months of age (Figures 3B, 3E). Likewise, increased FAO via ACC2 deletion did not affect gross cardiac morphology as both wall thickness and internal chamber dimensions were comparable in CON and ACC2H−/− up to 12 months of age (Figures 3C, 3D). Thus, chronic elevation of cardiac FAO is not harmful to overall cardiac performance.
To determine efficiency of ACC2H−/− hearts, we simultaneously measured myocardial oxygen consumption (MVO2) and contractile function in isolated Langendorff-perfused hearts (Figure 4A, 4B). MVO2 and contractile performance were similar in CON and ACC2H−/− when hearts were perfused with a buffer consisting of glucose and pyruvate as the substrates (Figure 4A). When the perfusate was switched to a buffer consisting of glucose, FA, and lactate, both MVO2 and contractile function, assessed by rate pressure product (RPP), increased slightly in ACC2H−/− although this was not statistically significant. The oxygen efficiency, estimated by MVO2/RPP, was not different between the groups (Figure 4B).
To determine whether altered substrate utilization observed in ACC2H−/− hearts affected ATP production and myocardial energetics, we measured dynamic changes of high energy phosphate content in isolated hearts during a high workload challenge with 31P NMR spectroscopy. Under normal workload conditions, the concentration of the energy reserve compound, phosphocreatine ([PCr]), was slightly higher in ACC2H−/− (p<0.05), with insignificant differences of inorganic phosphate ([Pi]) or ATP concentrations between the two genotypes (Figures 4C, 4E, 4G). At the end of the high workload challenge, [PCr] tended to be higher in ACC2H−/− (P = 0.08). While [ATP] was maintained in ACC2H−/−, this measure decreased ~10% in CON (P= 0.10; Figures 4C, 4E). Although [Pi] was significantly increased in both groups at the end of the high workload challenge, it was lower in ACC2H−/− (p<0.05, Figure 4G). Assessment of cardiac function showed that left ventricular developed pressure (LVDevP) and heart rate (HR) were similar in ACC2H−/− and CON under normal workload (Figures 4D, 4F, 4H). Both ACC2H−/− and CON had equivalent increases in LVDevP at similar HR in response to a high workload challenge (Figures 4D, 4F). However, EDP increased significantly in CON suggesting an impairment of relaxation during the high workload challenge, which was not observed in ACC2H−/− (Figure 4H). These observations suggest that elevated FAO provides equivalent if not slightly advantageous energetic support to cardiac function in ACC2H−/− mice.
To test the metabolic and functional response of cardiac-specific ACC2 deletion during the development of pressure-overload hypertrophy, CON and ACC2H−/− mice underwent sham or transverse aortic constriction (TAC) surgery. Acute mortality, defined as any death within 48hrs post-surgery, was not significantly different in CON-TAC or ACC2H−/−-TAC (26.3% vs. 26.2%). No deaths occurred as a result of sham surgery. To confirm that the pressure-overload achieved during surgery was comparable in both groups, pressure gradient measures were obtained using pulsed wave (PW) Doppler 24-hours after surgery in a cohort of hearts. No significant differences were noted between CON-TAC and ACC2H−/−-TAC (Online Figure I).
In pathological cardiac hypertrophy, FAO is reduced with increased reliance on carbohydrate sources for energy production. However, when ACC2H−/− were exposed to 8 weeks of pressure overload by (TAC), glucose and FAO remained similar to sham hearts (Figure 5A). Conversely, CON hearts undergoing 8 weeks of TAC had a significant decrease in FAO as compared to CON sham animals (Figure 5A), accompanied by an increase in glucose oxidation (Figure 5A). Glycolytic activity was assessed from the 13C enrichment patterns of alanine and lactate, derived from the glycolytic end product pyruvate, in extracts from isolated hearts perfused with 13C-labeled glucose. CON-TAC hearts had significant increases of both 13C-alanine and 13C-lactate, which was abrogated in ACC2H−/−-TAC (Figure 5B,5C). Furthermore, isotopomer analysis in extracts from hearts perfused with 13C labeled fatty acids and glucose revealed an ~2-fold increase in anaplerosis in CON-TAC, which was also not observed in ACC2H−/−-TAC (Figure 5D). All together, these data indicate that preservation of FAO prevents the metabolic shift towards increased reliance on glycolysis and anaplerosis in hypertrophied hearts.
We performed global metabolite profiling in CON and ACC2H−/− hearts to explore potential changes in the metabolic network in response to chronic elevation of FAO under normal and pressure overload conditions. Heart extracts were analyzed by GC × GC-TOFMS and chromatographic data were processed by the Fisher ratio (F-ratio) algorithm.34 Two dimensional (2D) Sum of F-ratio plots were produced to compare the four groups (Online Figure II) and over 800 hits were sorted in descending order and plotted against a null-distribution (Online Figure III). With this approach, we observed minimal effects of TAC and moderate effects of ACC2 deletion on the global metabolite profile. Identification of the top 15 hits yielded relatively few known metabolites. PARAFAC analysis did not show significant differences in nearly half of the hits (Online Tables III–VI). These findings suggest that there was no global cardiac metabolite shift during pressure-overload hypertrophy and/or with increased FAO as a result of ACC2 deletion.
We also incorporated a targeted approach specifically examining metabolites involved in FA, glucose, and amino acid (AA) metabolism as well as TCA cycle intermediates. We found significant decreases in a number of AAs and several metabolites in glucose metabolism with minor changes in TCA cycle intermediates in ACC2H−/−-sham hearts (Figure 5E), suggestive of decreased glucose reliance and increased AA consumption in ACC2H−/− hearts. These results propose that relative amounts of metabolites are maintained during the early development of pressure-overload hypertrophy despite a shift towards increased glucose utilization. Furthermore, adaptations to increased FAO in ACC2H−/− likely include reduced glucose uptake and utilization with increased consumption of AAs, which may contribute to the resistance to metabolic remodeling during pathological hypertrophy.
A hallmark of pathological cardiac hypertrophy is impaired myocardial energetics as reflected by significant decreases in the [PCr] despite minimal changes in [ATP] 35,36. In ACC2H−/−-TAC hearts, PCr/ATP ratios were maintained while a significant decrease in PCr/ATP was observed in CON-TAC relative to sham controls (Figure 6A). Cardiac function, assessed by the rate pressure product (RPP) in isolated perfused hearts, or assessed in-vivo via echocardiography, was sustained in ACC2H−/−-TAC hearts while CON-TAC showed a significant decline 8 weeks post-surgery (Figure 6B, 6C, Online Table VII).
To investigate the development of hypertrophy from pressure-overload, hearts from CON and ACC2H−/− mice were harvested 4 weeks post TAC. Both CON and ACC2H−/− hearts had a significant increase in brain natriuretic peptide (BNP) mRNA levels compared to sham counterparts (Figure 7A). However, the relative increase of BNP in ACC2H−/−-TAC hearts was approximately 50% less than CON-TAC. More importantly, cardiac hypertrophy, assessed by the heart weight to tibial length ratio (HW:TL), was ~50% less in ACC2H−/−-TAC compared to CON-TAC (Figure 7B). The attenuation of pathological hypertrophy was further evidenced by attenuation in myocyte cross sectional area and decreased fibrosis in ACC2H−/−-TAC versus CON-TAC (Figure 7C–7F). This pattern remained after 8 weeks of pressure-overload as molecular markers of hypertrophy, BNP and atrial natriuretic peptide (ANP), as well as a physical marker of hypertrophy, HW:TL, were significantly elevated in CON-TAC hearts with attenuation in ACC2H−/−-TAC (Online Figure IV). Consistent with our findings of decreased energetics, expression of genes associated with mitochondrial biogenesis (PGC1α) and mitochondrial ATP synthase activity (ATP5b) were significantly downregulated in CON-TAC, but not in ACC2H−/−-TAC (Online Figure IV), which was not associated with changes in mitochondrial volume as estimated by citrate synthase activity (Online Figure IV).
The current study addresses several key issues in cardiac metabolism. First, it is generally noted that enhanced FAO is not well tolerated in cardiac tissues, especially in conditions with elevated lipid delivery. However, we demonstrate that the heart is very capable of sustaining chronic increases of FAO when lipid supply is maintained. Our results show that it is unlikely that FAO per se but rather, the balance of lipid supply and oxidation is important. Second, despite strong evidence suggesting metabolic derangements that occur during the development of cardiac hypertrophy contribute to the transition to failure; prevention of metabolic reprogramming in pathological hypertrophy has not been accomplished. Here, we show that facilitating long-chain FA entry into the mitochondria via the ACC2/malonyl CoA/CPT1 mechanism will sustain the inherent cardiac metabolic profile during the development of pathological hypertrophy. Lastly, the question of whether mechanical dysfunction is a cause or consequence of metabolism has been debated. Our results suggest that altered cardiac metabolism leads to impaired function and energetics and that maintenance of cardiac metabolism preserves these measures during the early development of mild pressure-overload hypertrophy.
Conditions of chronic increased lipid delivery, as in obesity and diabetes, have been linked to dysfunction in liver, skeletal muscle, and heart.12,30,37 In the heart, increased FAO observed during reperfusion following ischemia formulated the basis that elevated FAO was the cause of poor functional recovery.38 Studies in PPARα transgenic mice13 or PPARα agonism in hypertrophied hearts8,9 suggested that enhanced FAO was harmful to the myocardium. However, PPARα has widespread effects on lipid metabolism, including fatty acid uptake, which may lead to an imbalance between FA uptake and oxidation. In skeletal muscle, increasing FAO via high fat feeding was associated with reduced TCA cycle intermediates (TCAI) combined with elevations in acylcarnitine species, suggesting that a portion of FAO is incomplete and contributes to insulin resistance.30 However, depletion of TCAI during high fat feeding was not seen in the heart,39 suggesting that chronic elevations in cardiac FAO are not detrimental. Our present data support this as increasing cardiac FAO via reductions in malonyl CoA is well tolerated without adverse effects on cardiac function in mice up to one year of age. A similar finding has been reported recently as mice with overexpression of PDK4 had increased FAO with normal cardiac function.40 Overall, these data suggest that mitochondrial capacity for FAO in heart tissue is more robust than other tissues and high FAO in the face of unchanged FA delivery is not a culprit for mechanical and mitochondrial dysfunction.
Targeted metabolic profiling revealed significant differences in glucose and amino acid metabolism with maintained TCA cycle intermediates between CON and ACC2H−/− hearts that remained during the development of pathological hypertrophy. The overall decreased presence of metabolites related to glucose metabolism in ACC2H−/− hearts likely represents an overall inhibition of glucose metabolism, including glucose uptake, which results from increased FAO as proposed by the Randle or “glucose-fatty acid cyle”.41 As a result of enhanced FAO, elevations in the [acetyl-CoA]/[CoA] and [NADH]/[NAD+] ratios lead to inhibition of pyruvate dehydrogenase (PDH) activity. Concurrently, accumulation of cytosolic citrate inhibits glycolysis via phosphofructokinase (PFK) in addition to glucose uptake.42–44 One potential negative outcome of increasing FAO is the production of acetyl CoA without the additional production of TCA cycle intermediates (TCAI) such that occurs with pyruvate.45 In ACC2H−/− hearts sustained TCAI was associated with increased consumption of glucogenic amino acids which can supply TCAI. The increased amino acid catabolism becomes particularly salient in light of the demonstration that disruption of branched chain amino acids (BCAA) catabolism through loss of protein phosphatase 2Cm (PP2Cm) was associated with mitochondrial dysfunction and heart failure.46–48 Therefore, increases in the contribution of amino acids to metabolism, especially from BCAAs, may play a significant role in cardiac metabolism during pathological stress and deserves attention in future research plans.
An unexpected observation in ACC2H−/− mice exposed to pressure-overload was the attenuation of LV hypertrophy. Unfortunately, the exact mechanism by which increased FAO via ACC2 deletion leads to an attenuation of cardiac hypertrophy is not known at this time. Interestingly, high fat feeding corresponded with reduced cardiac hypertrophy although the exact mechanism was also not uncovered.49 We speculate that the maintained metabolism and energetics in ACC2H−/− presents an adaptive phenotype which resists pathological stress, leading to attenuation of hypertrophic growth. However, it is also possible, that increases in amino acid metabolism, including the BCAAs, could divert amino acid availability, thus, retarding the normal upregulation of protein synthesis seen in pathological hypertrophy.50
Decreased FAO and increased glycolysis has been repeatedly demonstrated in models of pathological cardiac hypertrophy.1–3 Increased anaplerosis has been recently identified in hypertrophied hearts as well contributing to the metabolic phenotype.4,5 Although glucose has the advantage of being more oxygen-efficient compared to FA, it is not a carbon-efficient as only 2/3 of the carbon in glucose is oxidized compared to the complete oxidation of FA. In this regard, reliance on carbohydrate metabolism likely represents an energy-deficient state which predisposes the hypertrophied myocardium to contractile dysfunction.6,7 We have previously shown that substantial increases of glucose entry via an insulin-independent mechanism was necessary to rescue impaired FAO and prevent heart failure in mice.24 However, since achieving such increases in glucose uptake is not physiologically feasible, focusing on maintaining the inherent cardiac metabolic profile (i.e., reliance on FA over glucose) during the development of cardiac hypertrophy may be a preferred approach. In the present study, since deletion of ACC2 in pathological hypertrophy maintained cardiac FAO, normalized the expected increases in glycolysis and anaplerosis, and led to preserved cardiac function and energetics, metabolic therapies tailored toward this outcome may be more effective in preventing the transition to heart failure.
Previous work demonstrated partial reduction of cardiac malonyl CoA content as a result of global ACC2 deletion.16,20 We made similar findings suggesting that total removal of cardiac malonyl CoA is not achievable by ACC2 deletion. In our model, several possibilities exist for this: 1) ACC1 is present and likely contributes to a portion of the remaining malonyl CoA pool; 2) αMHC-Cre recombinase does not result in total deletion of protein, so there is residual malonyl CoA formed from ACC2; and 3) since αMHC-Cre is myocyte-specific, malonyl CoA from non-myocytes is still present. Nevertheless, our results clearly show that cardiac-specific deletion of ACC2 is sufficient to lower malonyl CoA levels and significantly increase mitochondrial FAO.
In summary, modulation of mitochondrial FA entry through cardiac-specific deletion of ACC2 increases FAO without adverse effects of cardiac function in mice up to 1 year of age. Furthermore, ACC2 deletion is sufficient to prevent metabolic reprogramming during pressure-overload hypertrophy with preservation of cardiac function and energetics, representing an overall state of reduced pathological stress leading to an attenuation of hypertrophic growth. These data suggest that maintenance of the inherent metabolic profile of the heart and, as a result, maintaining mitochondrial oxidative metabolism without switching to glucose reliance is beneficial for optimal function and energetics during the development of pathological cardiac hypertrophy. Although our data clearly demonstrate the benefits of preventing metabolic remodeling in the early course of pathological hypertrophy, additional studies are needed to verify that maintaining cardiac FAO is likewise beneficial in heart failure or other cardiac pathologies.
We thank Dr. Ronglih Liao and Soeun Ngoy at the Animal Physiology Core, Brigham and Women’s Hospital (Boston, MA) for generating the mouse TAC models. We also thank Drs. Chris Liu and Lei Li from the Department of Physiology and Biophysics at University of Washington (Seattle, WA) for assistance with pressure gradient measurements across the aortic constriction.
SOURCES OF FUNDING
This work is supported by NIH grants HL059246, HL067970 (to RT), and HL096284 (to SK).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.