Aberrant cardiac metabolism has long been hypothesized to contribute to heart failure (HF). However, understanding of cardiac metabolism in HF remains incomplete, especially in humans. To fill this gap, we performed metabolomic and RNA sequencing analysis of 48 non-failing (NF) and 39 dilated cardiomyopathy (DCM) hearts. Among many differences, metabolites from one-carbon metabolism were significantly reduced, especially in methionine cycle; methionine (Met) and S-adenosylmethionine (SAM), a universal methyl donor of methylation. We next studied the mechanism leading to decreased Met/SAM in DCM hearts. We first quantitatively defined methionine cycle in the heart: studies in neonatal rat ventricular myocytes (NRVMs), and in vivo steady-state infusions with isotope tracing, showed that Met was mostly imported or degraded from protein rather than recycled from homocysteine (~10%), suggesting that a defect in recycling is unlikely to explain low Met/SAM in DCM hearts. We next explored the impact of low SAM on cardiac function by generating mice with cardiac specific deletion of SAM synthase. Constitutive KO mice were largely embryonic lethal, and the few survivors showed dramatic systolic dysfunction and died within 12 weeks. Inducing loss of SAM synthase in adults led to 50% mortality after 2 weeks, again with marked cardiac dysfunction. Finally, we explored potential mechanisms by which low SAM can drive heart failure. In human DCM, numerous methylation products were low, including most notably creatine, a SAM methylation product and essential metabolite for cardiac energetics. Creatine is not thought to be synthesized in the heart, but we found that, to our surprise, NRVMs synthesized ~50% of their creatine pool when external supply was low, and synthesis was dependent on SAM availability. Moreover, in vivo steady state infusions revealed that in intact animals, even in the absence of stressors, ~10% of creatine in the heart is synthesized locally. In summary, we find that: (1) the majority of Met in cardiomyocyte is imported or derived from protein degradation, with rapid turnover; (2) Met and its product SAM are reduced in human DCM; (3) loss of SAM in murine heart leads to profound heart failure; (4) creatine is surprisingly synthesized from SAM in cardiomyocytes, and its levels are reduced in human DCM. Together, these findings suggest that a defective Met cycle may contribute to low creatine and ultimately HF.

<p>Figure S1. Experimental Figure Scheme. Figure S2. ChIP-Seq quality assessment for different samples and histone marks. Figure S3. Comparable ChIP-Seq and ChIP-qRT-PCR detection of histone enrichment for different histone marks near three reference genes. Figure S4. Comparable expression of reference genes. Figure S5. H3K9ac histone mark enrichment distribution near transcription start sites. Figure S6. H3K27ac histone mark enrichment distribution near transcription start sites. Figure S7. H3K9me3 histone mark enrichment distribution near transcription start sites. Figure S8. Correlation between tissue specific H3K9ac histone peak enrichment and expression of nearby genes. Figure S9. Correlation between tissue specific H3K9me3 histone peak enrichment and expression of nearby genes. Figure S10. Correlation between tissue specific H3K4me3 histone peak enrichment and expression of nearby genes for six study samples and two PDX-parental tissues. Figure S11. Correlation between tissue specific H3K27ac histone peak enrichment and expression of nearby genes for six study samples and two PDX-parental tissues. Figure S12. Correlation between tissue specific H3K9ac histone peak enrichment and expression of nearby genes for six study samples and two PDX-parental tissues. Figure S13. Correlation between tissue specific H3K9me3 histone peak enrichment and expression of nearby genes for six study samples and two PDX-parental tissues. Figure S14. Scheme for integration of ChIP-Seq Specific Histone Peaks with Expression Variation Analysis (EVA). Figure S15. Correlation of H3K27ac-enriched with genes differentially regulated in mesenchymal and classical HPV+ HNSCC subtypes.</p>

SFigure 1. Distribution of ASEs present within each tumor. SFigure2. AKT3 splice variant expression by qRT-PCR in a panel of 20 cell lines. SFigure 3. Western blot analysis of canonical AKT1 pathway after transient knockdown of AKT3 splice variant. SFigure 4. Correlation of AKT3 splice variant with other PI3K/AKT pathway mutations in TCGA.

Recent studies suggest that metabolic dysregulation in patients with heart failure might contribute to myocardial contractile dysfunction. To understand the correlation between function and energy metabolism, we studied the impact of different fuel substrates on human nonfailing or failing cardiomyocytes. Consistent with the concept of metabolic flexibility, nonfailing myocytes exhibited excellent contractility in all fuels provided. However, impaired contractility was observed in failing myocytes when carbohydrates alone were used but was improved when additional substrates were added. This study demonstrates the functional significance of fuel utilization shifts in failing human cardiomyocytes.

<p>Table S1. Clinical data for patient-derived PDX1, PDX2, UPPP1, and UPPP2 samples used for ChIP-based analysis. Table S2. ChIP-DNA qRT-PCR primers-probe sequence Table S3. Sample-specific enrichment of H3K4me3 histone mark at 5''UTR of individual genes. Table S4. Sample-specific enrichment of H3K27ac histone mark at 5''UTR of individual genes. Table S5. Sample-specific enrichment of H3K9ac histone mark at 5''UTR of individual genes. Table S6. Sample-specific enrichment of H3K9me3 histone mark at 5''UTR of individual genes. Table S7. Gene set enrichment analysis of genes linked to H3K27ac-enrichment specific for tumor samples. Table S8. Gene set enrichment analysis of genes linked to H3K27ac-enrichment specific for normal samples. Table S9. Gene set enrichment analysis of genes linked to H3K4me3-enrichment specific for tumor samples. Table S10. Gene set enrichment analysis of genes linked to H3K4me3-enrichment specific for normal samples. Table S11. Correlation of H3K27ac-enriched with genes differentially regulated in HPV-KRT and HPV-IMU.</p>

Aberrant cardiac metabolism has long been hypothesized to contribute to heart failure (HF). However, understanding of cardiac metabolism in HF remains incomplete. To fill this gap, we performed metabolomic and RNA sequencing analysis of 48 non-failing (NF) and 39 dilated cardiomyopathy (DCM) hearts. Among many differences, metabolites from one-carbon metabolism were significantly reduced, especially methionine (Met) and s-adenosylmethionine (SAM), a universal methyl donor of methylation. We probed the mechanism behind decreased Met/SAM in DCM hearts. Studies in neonatal rat ventricular myocytes (NRVMs), and in vivo steady-state infusions with isotope tracing showed that Met was mostly imported, rather than recycled, suggesting a defective Met transport in DCM hearts. We next explored the impact of low SAM. DNA and histone methylation remained unchanged, but creatine level was significantly decreased. So, we tested potential creatine synthesis in the heart that was previously unclear. To our surprise, NRVMs synthesized ~50% of their creatine pool when external supply is low, and synthesis was dependent on SAM availability. In summary, we find that the majority of Met in cardiomyocyte is imported, with rapid turnover. We also find that cardiomyocytes synthesize creatine and increase creatine production under stress. Together, these findings suggest that defective Met import may contribute to low creatine and SAM levels and ultimately HF.

The heart is the most metabolically active organ in the body, sustaining a continuous and high flux of nutrient catabolism via oxidative phosphorylation. The nature and relative contribution of these fuels have been studied extensively for decades. By contrast, less attention has been placed on how intermediate metabolites generated from this catabolism affect intracellular signaling. Numerous metabolites, including intermediates of glycolysis and the tricarboxylic acid (TCA) cycle, nucleotides, amino acids, fatty acids and ketones, are increasingly appreciated to affect signaling in the heart, via various mechanisms ranging from protein–metabolite interactions to modifying epigenetic marks. We review here the current state of knowledge of intermediate metabolite signaling in the heart.

Designed primer sequences for wild type AKT3 and variant AKT3 for qRT-PCR analysis as well as custom siRNA sequences

Heart failure is marked by metabolic insufficiency, but detailed understanding of the underlying metabolic rewiring has been limited. By applying state-of-the-art mass spectrometry to a large pool of human hearts in end-stage heart failure, we unveil numerous metabolic aberrations in human heart failure.

The script for ChIP-Seq and RNA-Seq data integration

R code for splice variant analysis and outlier statistics for identification of significant splice variants

Cell line authentication for SCC 61

List of 109 validated junctions associated with splice variants with location, gene symbol and category.

Summary of sample cohort clinical data in tumors and normal samples including age, gender, race, smoking history, TNM stage, recurrence, follow up time, survival and HPV status

Pharmacologic activation of branched chain amino acid (BCAA) catabolism is protective in numerous models of heart failure (HF). How this protection occurs has remained unclear, although a causative block in cardiac BCAA oxidation has been proposed. We use here in vivo heavy isotope infusion studies to show that cardiac preference for BCAA oxidation increases, rather than decreases, in multiple models of HF. We use various genetic models to show that cardiac-specific activation of BCAA oxidation does not protect from HF, even though systemic activation of BCAA oxidation does. Lowering plasma and cardiac BCAAs by genetic means is also not sufficient to confer protection comparable to that conferred by pharmacologic activation of BCAA oxidation, suggesting alternative mechanisms of protection. Surprisingly, telemetry and invasive hemodynamic studies showed that pharmacological activation of BCAA catabolism lowers blood pressure, a well-established cardioprotective mechanism. The effects on blood pressure occurred independently of nitric oxide (NO), and reflected a vascular resistance to adrenergic constriction. Finally, mendelian randomization studies revealed that elevations in plasma BCAAs portend higher blood pressure in large human cohorts. Together, these data indicate that activation of BCAA oxidation lowers blood pressure and protects from heart failure independently of any direct effects on the heart itself.

Heart failure affects millions of people worldwide with mortality near 50% within five years. This disease is characterized by widespread cardiac and systemic metabolic changes, but a comprehensive evaluation of metabolism in failing human hearts is lacking. Here, we provide a comprehensive depiction of cardiac and systemic metabolic changes in 89 explanted failing and non-failing human hearts through integration of plasma and cardiac tissue metabolomics, genome-wide RNAseq, and proteomic data. The data confirm a profound bioenergetic defect in end-stage human heart failure and demonstrate extensive changes in metabolic homeostasis. The data indicate a substantial defect in fatty acid (FA) use in failing hearts, in particular unsaturated FAs. Reduction of FAs and acyl-carnitines in failing tissue in contrast to concomitant elevations in plasma suggest a defect in import of FAs into the cell, rather than a defect in FA oxidation. Intermediates of glycolysis, the pentose phosphate pathway, and glycogen synthesis are all similarly reduced, as is expression of GLUT1, indicating diminished glucose uptake. However, there was no significant change in tissue pyruvate content, suggesting an increase in lactate utilization. The data suggest increased flux of pyruvate into mitochondria, likely promoting pyruvate oxidation but not pyruvate carboxylation. Blunted anabolic pyruvate flux, in turn, likely leads to insufficient TCA cycle intermediates. Ketone levels were increased in both failing tissue and plasma, as previously reported. The phospholipid content of failing human hearts is greatly increased in both failing tissue and plasma. Nucleotide synthesis pathways also appear to be reprogrammed, with a notable decrease in adenosine metabolism, specifically. Together, these data indicate widespread change in the local cardiac and greater systemic metabolic landscape in severe human heart failure.

Heart failure (HF) is a leading cause of mortality. Failing hearts undergo profound metabolic changes, but a comprehensive evaluation in humans is lacking. We integrate plasma and cardiac tissue metabolomics of 678 metabolites, genome-wide RNA-sequencing, and proteomic studies to examine metabolic status in 87 explanted human hearts from 39 patients with end-stage HF compared with 48 nonfailing donors. We confirm bioenergetic defects in human HF and reveal selective depletion of adenylate purines required for maintaining ATP levels. We observe substantial reductions in fatty acids and acylcarnitines in failing tissue, despite plasma elevations, suggesting defective import of fatty acids into cardiomyocytes. Glucose levels, in contrast, are elevated. Pyruvate dehydrogenase, which gates carbohydrate oxidation, is de-repressed, allowing increased lactate and pyruvate burning. Tricarboxylic acid cycle intermediates are significantly reduced. Finally, bioactive lipids are profoundly reprogrammed, with marked reductions in ceramides and elevations in lysoglycerophospholipids. These data unveil profound metabolic abnormalities in human failing hearts. Arany and colleagues integrate metabolomics data, tissue RNA-sequencing data and proteomics data from explanted hearts of patients with end-stage heart failure to provide a comprehensive ‘multi-omic’ evaluation of the metabolic pathways affected by heart failure.

Truncating variations in the gene coding for titin (TTNtv) have been known to cause dilated cardiomyopathy for nearly 20 years. Efforts to detect direct evidence of either haploinsufficiency or dominant negative mechanisms have thus far failed, leaving the mechanism open to controversy. By analyzing a collection of 184 post-transplant human hearts, 22 of which bear TTNtv’s, we show evidence supporting both haploinsufficient (lack of sufficient full length titin to maintain normal cardiomyocyte contractility) and dominant-negative (toxic gain of function due to truncated titin) mechanisms. Using allele specific proteomics as well as epitope specific agarose gel immunoblotting we show that TTNtv are present in human myocardium at the expected molecular weight and bear only the epitopes expected to be present in TTNtv protein. TTNtv associate with the sarcomere bearing insoluble fraction of human myocardium but are more weakly attached to the sarcomere than full length titin, consistent with their lack of thick filament and M-line attachment sites. We further show that DCM hearts bearing TTNtv have less full length titin than non-TTNtv bearing DCM hearts, by both total protein and in ratio to sarcomeric proteins, indicating TTN haploinsufficiency is also present in TTNtv hearts. This unambiguous detection of TTNtv protein in the myocardium of DCM combined with a reduction in full length titin supports a combined dominant negative and haploinsufficient mechanism of pathogenesis of TTNtv induced DCM.