Myocardial Substrate Changes in Advanced Ischaemic and Advanced Dilated Human Heart Failure

There has been renewed interest in cardiac metabolism at least partly driven by the dramatic improvements in heart failure (HF) outcomes in recent sodium–glucose co-transporter 2 (SGLT2) inhibitor trials.1 It has been proposed that these agents provide ‘thrifty’ substrates that the failing heart can leverage for more efficient energy supply; however, this process is poorly understood. Most prior reports have used circulating metabolites or lipids to infer conclusions regarding human cardiac substrate use.2 However, it is unclear if these perturbations, remote from the heart, adequately reflect actual myocardial substrate metabolism. Therefore, we investigated changes across multiple metabolic pathways in human left ventricular myocardial tissue from 26 end-stage HF hearts: ischaemic cardiomyopathy (ICM) (n = 14) and non-ischaemic dilated cardiomyopathy (DCM) (n = 12), compared with matched healthy donor hearts (n = 17). Two targeted platforms were used to measure 210 known metabolites on our platforms at Sydney Mass Spectrometry: hydrophilic interaction chromatography (HILIC)–tandem mass spectrometry with positive electrospray ionisation, and HILIC chromatography/amide stationary phase–tandem mass spectrometry and negative electrospray ionisation.3 These methods incorporate a broad range of metabolites including amino acids, acylcarnitines, ketone bodies, energetics and central carbon metabolism. These methods are highly reproducible, with an average coefficient of variation of 3.85% across all metabolites.4 Nevertheless, two independent mass spectrometry runs were performed for each method, which were combined and analysed. We focused on the 40 metabolites reported in the table that were present in all tissue samples in both independent LC–MS runs, and that passed internal quality control assessment. Therefore, we did not adjust for multiple comparison. Data were analysed using Wilcoxon rank-sum test using SPSS (SPSS 20, Chicago, IL, USA). Metabolites and proteins with a P-value < 0.05 were considered to be statistically significant. Mean age in each group was as follows (mean ± standard deviation): age ICM 59 ± 8, DCM 47 ± 12, donor 45 ± 15 years. Left ventricular HF samples (ICM and DCM) were procured from hearts of New York Heart Association class III–IV patients with left ventricular ejection fraction (LVEF) <35% at the time of heart transplantation. All patients, regardless of the aetiology of HF, received guideline-directed medical therapy pre-transplant, including angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker, β-blocker and spironolactone. No patients were on mechanical assist devices at the time of transplantation. Age-matched donor hearts were procured, but not used for heart transplantation (primarily for logistical reasons). They did not have any known cardiovascular risk factors and all hearts underwent formal pathology examination to ensure they were histologically normal, donor hearts did not show histological pathology. Both donor and recipient hearts were procured after preceding cardioplegia. All hearts (donor and HF) were immediately processed and stored in liquid nitrogen according to published protocols5 (Sydney Heart Bank HREC # 2016/923). Fibrotic tissue was avoided when processing samples for metabolomics. Subsequent analysis showed the intra-group variability in metabolite levels to be low, suggesting that myocardial tissue composition was not a significant issue influencing metabolite results. Samples were crushed in liquid nitrogen to powder and exactly 50 mg tissue used for each sample. We found significant changes in ICM and DCM myocardium in many pathways. Uridine and its metabolites 2-deoxycytidine, cytidine, and cytosine were significantly increased in both ICM and DCM. Uridine is reported as being pathological in HF: in vivo administration of uridine triphosphate induced myocardial fibrosis, and in vitro uridine diphosphate initiated fibrosis upon mechanical stretch via its P2Y6 receptor.6 Branched-chain amino acids (BCAAs), including valine, leucine, and isoleucine, were significantly elevated in both ICM and DCM. Mechanical stress up-regulates catabolism of BCAAs, causing cardiac dysfunction either through toxic intermediates – branched-chain α-keto-acids – or through direct activation of the protein kinase mTOR.7 Methionine, homocysteine, and derivatives were also significantly changed. The homocysteine pathway has been widely implicated in cardiovascular disease, particularly in the promotion of atherosclerosis via endothelial dysfunction, low-density lipoprotein oxidation, and inflammation. However, this is the first report, to our knowledge, that demonstrates decreased homocysteine in HF hearts. There was increased lipid metabolism in HF myocardium. Short-chain acylcarnitines isovalerylcarnitine, butyrylcarnitine, and 2-methylbutyrylcarnitine were significantly decreased in ICM and DCM compared with donor hearts. These data are consistent with previous reports of increased utilisation of short-chain acylcarnitines in advanced HF and a low myocardial:plasma ratio.2 Ketone bodies were also significantly changed. Acetoacetate was significantly decreased in DCM compared with donor, whereas β-hydroxybutryrate was significantly increased in DCM compared with donor hearts. However, there was no consistent change in ketone bodies across both types of HF, this finding is at odds with reports suggesting that the failing heart is dependent on ketone bodies as its major substrate. Interestingly, recent work examining mechanisms underpinning benefits of SGLT2 inhibitor empagliflozin in the hearts of db/db mice reported increased glucose and fatty acid oxidation, but no change in ketone body oxidation.8 Deoxyribose-5-phosphate, a pentose phosphate pathway (PPP) intermediate, was significantly elevated in both ICM and DCM, aligned with reported elevation of PPP intermediates post-myocardial infarction. Consistent with previous reports of reduced metabolic flexibility and changes in glycolysis in advanced HF, we found glycolytic intermediates glucose-6-phosphate and fructose-6-phosphate to be significantly reduced in ICM myocardium compared with donor hearts. These changes and other metabolomic changes are summarised in Figure 1. In this letter, for the first time, we provide a cross-sectional report on human myocardial substrate perturbations common to HF of different aetiologies, as compared with matched, healthy donor hearts. The findings in the current study were limited to those with end-stage, as opposed to earlier-stage, HF. Indeed, a spectrum of disease severity and associated metabolite changes is likely to exist, and should be more rigorously explored in animal models in relation to time course and severity of HF development. Our work significantly informs recent work using circulating metabolites and previous murine models of HF. Furthermore, this work implicates many of the metabolic pathways found to be altered by SGLT2 inhibition. Conflict of interest: none declared. Glycine Alanine Serine Proline Valine Threonine Cysteine Isoleucine_Leucine Isoleucine Leucine Aspartate Lysine Glutamine Glutamate Methionine Histidine Glucose Phenylalanine Arginine Tyrosine Methionine sulfone (MetSul) Methylsulphone (MES) Tryptophan Asparagine Valine-d8 Phenylalanine-d8 1-Methylhistamine 2′-Deoxyadenosine 2′-Deoxycytidine 3-Hydroxykynurenine (3-HK) 5-Hydroxyindoleacetic acid (5-HIAA) Adenosine Asymmetric dimethylarginine (ADMA) α-Keto-δ-(NG,NG-dimethylguanidino) valeric acid (DMGV) Anserine β-Alanine Betaine Butyrlcarnitine Cyclic adenosine monophosphate (cAMP) Carnosine Choline Citrulline Cotinine Creatine Cystamine Cysteamine Cytidine Cytosine γ-Aminobutyric acid (GABA.1) GABA.2 Glycerol Histamine Kynurenic acid Ornithine Phosphocholine Phosphoethanolamine Serotonin Spermine Taurine Thiamine Thyroxine trans-Hydroxyproline (trans-HYP) Triidothyronine Trimethylamine N-oxide (TMAO) Thymidine Uridine 3-Deaazadenosine 3-Indolepropionic acid (3-IPA) β-Aminoisobutyric acid (BAIBA) Acetylcarnitine Acetylcholine Carnitine l-Homoserine N-Acetylglutamine NG-Monomethyl-l-arginine (L-NMMA) Purine Pyridoxine Riboflavin Anandamide 2-Arachidonyl glycerol Arachidonic acid Aminoadipic acid (AAD) Colchicine Homocysteine Acetoacetic acid Glyoxylic acid 5-Aminolevulinic acid a-Ketoisocaproic acid.1 a-Ketoisocaproic acid.2 Isovalerylcarnitine.1 Isovalerylcarnitine.2 a-Keto-β-methylvaleric acid.1 a-Keto-β-methylvaleric acid.2 2-Methylbutyrlcarnitine.1 2-Methylbutyrlcarnitine.2 2-Methylbutyrlcarnitine.3 α-Ketoisovaleric acid (KIV) Propylcarnitine 2-Hydroxy-2-methylbutyrate (2H2MB) 2-Hydroxyglutarate (2-HG) 2-Isopropylmalate (2-IPM) 2-Ketohexanoate (2-KH) 2-Methylacetoacetate (2-MAA) 2-Oxobutanoate (2-OB) 3-Methyl-2-oxobutyrate (3M2OB) 2-Oxodipate (2-OD) 3-Hydroxy-3-methylglutaryl coenzyme A 3-Hydroxybutyrate (3-HB) 3-Hydroxykynurenine (3-HK) 3-Phosphoglycerate (3-PG) 4-Hydroxyphenylacetate (4-HPA) 5-Hydroxyindoleacetate (5-HIAA) Acetylphosphate (ACP) Aconitate ADP.1 ADP.2 Allantoin 2-Aminoadipate (AAD) Ascorbate ATP.1 ATP.2 Cholesteryl sulfate (CholSO4) Cyclic adenosine monophosphate (cAMP) Cholate Citraconate Deoxycholate (DCA) Deoxyribose-phosphate (DRP) Fructose 1,6-bisphosphate (Fruc-1,6-bP) Fructose 6-phosphate Oxidized glutathione (GSSG) Reduced glutathione (GSSH) Glyceraldehyde Guanidoacetate (GAA) Homocysteate (HCA) Homogentisate Hydroxyisocaproate (HICA) Isocitrate Kynurenate Kynurenine Malate Fumarate Malonyl-coenzyme A Mevalonate (MEV) N-Acetyl-l-alanine Nicotinate p-Aminobenzoate (p-AB) Phenyllactic Phenylpyruvate Phosphoenolpyruvate (PEP) Pyroglutamate Pyrophosphate Pyruvate Saccharopine S-Adenosyl-l-homocysteine Taurine Taurodeoxycholate (TDCA) Thiamine pyrophosphate (TPP) Thymidine Uracil Uridine Xanthosine Xanthurenate (XAN) Acetyl-coenzyme A Adenosine monophosphate (AMP.1) Adenosine monophosphate (AMP.2) Anthranilate (AA) Arginosuccinate Biotin Erythrose 4-phosphate (E4P.1) E4P.2 Glyceraldehyde 3-phosphate (G3P) Glutarylcarnitine (C5-DC) Inosine Malonate N-Carbomyl-aspartate Niacinamide Oxaloacetate Phytonadione Prephenate Quercetin Riboflavin 5′-monophosphate (R5'MP.1) R5′MP.2 Tropisetron UDP-N-acetyl-glucosamine Ureidopropionate Uridine 5′-diphosphate (U5′dP) Phenylalanine-d8 Thymine-d4 Glucose 6-phosphate Succinate Lactate Camphorsulfonate (CSA) NADH Glucose Nicotinamide adenine dinucleotide (NAD) Reduced nicotinamide adenine dinucleotide phosphate (NADPH) Nicotinamide adenine dinucleotide phosphate (NADP.1) NADP.2 Citrate-D4 Citrate 3-Hydroxyanthranilate (3-HAA)