Cardiac Dysfunction and Metabolic Inflexibility in a Mouse Model Of

Home , OGDH, Pyruvate dehydrogenase, Pyruvate synthase

Diabetes Volume 67, June 2018 1057

Cardiac Dysfunction and Metabolic Inﬂexibility in a Mouse Model of Diabetes Without Dyslipidemia

Maria Rohm,1 Dragana Savic,2 Vicky Ball,2 M. Kate Curtis,2 Sarah Bonham,3 Roman Fischer,3 Nathalie Legrave,4 James I. MacRae,4 Damian J. Tyler,2 and Frances M. Ashcroft1

Diabetes 2018;67:1057–1067 | https://doi.org/10.2337/db17-1195

Diabetes is a well-established risk factor for heart disease, and heart failure (1,2). Alterations in cardiac metabolism, leading to impaired cardiac function and a metabolic switch including changes in substrate utilization and mitochondrial toward fatty acid usage. In this study, we investigated if dysfunction, contribute to impaired heart function in di- hyperglycemia/hypoinsulinemia in the absence of dyslipi- abetes (3–6). The heart uses both glucose and free fatty acids demia is sufficient to drive these changes and if they can be (FFA) as fuel, with FFA accounting for 60–70% of energy reversed by restoring euglycemia. Using the bV59M mouse generation under normal conditions (7). FFA are metabo- model, in which diabetes can be rapidly induced and re- lized to acetyl-CoA by b-oxidation, and if FFA uptake exceeds versed, we show that stroke volume and cardiac output b-oxidation, as in obesity, this leads to accumulation of lipids were reduced within 2 weeks of diabetes induction. Flux and lipid metabolites, eventually causing lipotoxicity. By METABOLISM through pyruvate dehydrogenase was decreased, as mea- contrast, chronic hyperglycemia produces glucotoxicity, sured in vivo by hyperpolarized [1-13C]pyruvate MRS. causing the formation of reactive oxygen species and ad- Metabolomics showed accumulation of pyruvate, lactate, vanced glycation end products (6). Impaired cardiac function alanine, tricarboxyclic acid cycle metabolites, and branched- fi chain amino acids. Myristic and palmitoleic acid were de- is common to all types of diabetes and not con ned to creased. Proteomics revealed proteins involved in fatty acid individuals with diabetes with obesity or dyslipidemia. How- metabolism were increased, whereas those involved in glu- ever, the relative contributions of lipotoxicity and glucotox- cose metabolism decreased. Western blotting showed en- icity to cardiac dysfunction in diabetes remain unclear. hanced pyruvate dehydrogenase kinase 4 (PDK4) and Most studies to date have focused on dyslipidemia and uncoupling protein 3 (UCP3) expression. Elevated PDK4 shown that elevated serum FFA drive metabolic alterations and UCP3 and reduced pyruvate usage were present in the heart (6,8–10). The role of hyperglycemia has been less 24 h after diabetes induction. The observed effects were well investigated. However, the fact that the hypertrophic independent of dyslipidemia, as mice showed no evidence of cardiomyopathy and left ventricular dysfunction in a mouse elevated serum triglycerides or lipid accumulation in periph- model of lipodystrophy were ameliorated by lowering blood eral organs (including the heart). The effects of diabetes were glucose levels (11) suggests hyperglycemia may contribute to reversible, as glibenclamide therapy restored euglycemia, impaired cardiac function in diabetes. Furthermore, non- cardiac metabolism and function, and PDK4/UCP3 levels. obese patients with diabetes (including type 1 or monogenic diabetes) may also develop cardiac complications (12,13), Diabetes is an increasing health burden worldwide and supporting the idea that hyperglycemia alone is sufficient to a major risk factor for developing cardiovascular disease cause cardiac disease. Likewise, in a large study of individuals

1Department of Physiology, Anatomy and Genetics and OXION, University of M.R. is currently afﬁliated with the Institute for Diabetes and Cancer (IDC), Oxford, Oxford, U.K. Helmholtz Center Munich, Neuherberg, Germany; Joint Heidelberg-IDC 2Cardiac Metabolism Research Group, Department of Physiology, Anatomy and Translational Diabetes Program, Heidelberg University Hospital, Heidelberg, Genetics, University of Oxford, Oxford, U.K. Germany; Molecular Metabolic Control, Medical Faculty, Technical University 3Discovery Proteomics Facility, Target Discovery Institute, University of Oxford, of Munich, Munich, Germany; German Center for Diabetes Research, Neuherberg, Oxford, U.K. Germany. 4 The Francis Crick Institute, London, U.K. © 2018 by the American Diabetes Association. Readers may use this article as Corresponding author: Frances M. Ashcroft, [email protected]. long as the work is properly cited, the use is educational and not for proﬁt, and the Received 3 October 2017 and accepted 12 March 2018. work is not altered. More information is available at http://www.diabetesjournals .org/content/license. This article contains Supplementary Data online at http://diabetes .diabetesjournals.org/lookup/suppl/doi:10.2337/db17-1195/-/DC1. 1058 Cardiac Dysfunction in a Eulipidemic Model of Diabetes Diabetes Volume 67, June 2018

with type 2 diabetes (T2D), a 1% increase in HbA1c was (Innovative Research of America) and under 2% isoflurane associated with an increased risk of heart failure, indepen- anesthesia. A schematic overview of the mouse experiments dent of obesity, suggesting hyperglycemia is an independent is given in Supplementary Fig. 1D. For tissue analysis, mice risk factor for cardiac dysfunction (14). were culled by cervical dislocation, and hearts were collected, In this study, we used an inducible mouse model of snap-frozen in liquid nitrogen, pulverized, and stored at human neonatal diabetes to explore the effects of chronic 280°C for later analysis. hyperglycemia/hypoinsulinemia in the absence of dyslipidemia (15,16) (bV59M mice). Our mouse expresses an acti- Blood Metabolites Blood glucose levels were measured from the tail vein using vating mutation (Kir6.2-V59M) in the KATP channel specifically in pancreatic b-cells, which prevents glucose- a FreeStyle Lite device and FreeStyle Lite test strips (Abbott stimulated insulin secretion. It has several advantages Laboratories). Serum was obtained by incubating whole over other mouse models of diabetes. First, diabetes is bloodonicefor30minfollowedbycentrifugationat caused by a b-cell–specific genetic defect rather than a toxin 3,000g and 4°C for 30 min. It was then snap-frozen in liquid (e.g., streptozotocin) that may have deleterious effects in nitrogen for later analysis. Serum glucose was measured other tissues. Second, diabetes is not associated with obesity using a glucose (HK) kit (Sigma-Aldrich). Serum triglyceride or insulin resistance. Third, diabetes can be induced in adult (ab65336; Abcam), FFA (ab65341; Abcam), cholesterol (ab65390; Abcam), and insulin (10-1247-01; Mercodia) lev- life, precluding compensatory developmental changes. els were measured using the indicated kits. Fourth, diabetes is rapidly reversible by treatment with sulphonylurea drugs (e.g., glibenclamide), which close the Cine MRI open KATP channels (13). It should be recognized, however, Mice were imaged on an 11.7T MRI instrument (Bruker) as that this mouse is both hyperglycemic and hypoinsulinemic, previously described (17). Eight to 10 short-axis slices (slice as is the case in both type 1 diabetes and nonobese T2D. thickness, 1.0 mm; matrix size, 256 3 256; field of view, In this study, we show that plasma hyperglycemia (in 25.6 3 25.6 mm; echo time/repetition time, 1.43/4.6 ms; conjunction with hypoinsulinemia) is a major driver of flip angle, 17.5°; and number of averages, 4) were acquired impaired cardiac metabolism and function. Gene induction with a cine fast low-angle shot sequence (18). Left ventric- caused a substrate switch from glucose oxidation toward ular volumes were derived using the freehand draw function lactate production and increased fatty acid (FA) metabo- in ImageJ (National Institutes of Health). For each heart, left lism. This led to a reduced cardiac output and stroke volume. ventricular mass, ejection fraction, stroke volume, and car- Restoration of euglycemia by glibenclamide completely re- diac output were calculated. stored cardiac metabolism and function after 2 weeks of diabetes. The time-dependent deterioration of cardiac func- Hyperpolarized MRS tion correlated with reduced glucose flux through pyruvate Experiments were performed between 7 and 11 A.M. when 13 dehydrogenase (PDH) and increased expression of PDH mice were in the fed state. A total of 40 mg [1- C]pyruvate kinase 4 (PDK4) and uncoupling protein 3 (UCP3). (Sigma-Aldrich) doped with 15 mmol/L trityl radical (OXO63; GE Healthcare) and 3 mL Dotarem (1:50 dilution; Guerbet) – RESEARCH DESIGN AND METHODS was hyperpolarized in a prototype polarizer, with 20 30 min of microwave irradiation (19). The sample was subsequently Animals dissolved in a pressurized and heated alkaline solution, Animal studies were conducted in accordance with the U.K. containing 2.4 g/L sodium hydroxide and 100 mg/L EDTA Animals (Scientific Procedures) Act (1986) and local ethical dipotassium salt (Sigma-Aldrich), to yield a solution guidelines (Medical Research Council’sResponsibilityinthe of 80 mmol/L hyperpolarized sodium [1-13C]pyruvate Use of Animals in Medical Research, 1993). Mice hemi- with a polarization of 30%. A total of 200 mLwasinjected zygously expressing an inducible Kir6.2-V59M transgene 13 b b over 10 s via the tail vein. C MR pulse-acquire cardiac selectively in pancreatic -cells ( V59M mice) were gener- spectra were acquired over 60 s following injection of hyper- ated and transgene expression induced by tamoxifen in- polarized [1-13C]pyruvate (repetition time, 1 s; excitation flip b jection, as described (15). In some experiments, V59M mice angle, 15°; sweep width, 13,021 Hz; acquired points, 2,048; were used before gene induction as their own controls. In and frequency centered on the C1 pyruvate resonance) (20). other experiments, tamoxifen-injected or uninjected wild- The 13C label from pyruvate and its metabolic products was fl type, RIPII-Cre-ER, and oxed Kir6.2-V59M gene littermates summed over 30 s from the first appearance of pyruvate were used as controls. Mice were maintained on a 12-h and fitted with the AMARES algorithm in jMRUI (21). 6 light/dark cycle at 21 2°C with an unrestricted diet (63% Data are reported as the ratio of metabolic product to the carbohydrate, 23% protein, and 4% fat; RM3, Special Diet [1-13C]pyruvate signal to normalize for differences in polar- Services). ization and delivery. Body weight and blood glucose levels were monitored routinely. For glibenclamide treatment, after 2 weeks of Tissue Triglyceride Measurement diabetes, mice were subcutaneously implanted with two Cardiac triglycerides were isolated using the chloroform/ pellets, each releasing 25 mg glibenclamide over 21 days methanol extraction (adapted from Folch et al. [22]). In diabetes.diabetesjournals.org Rohm and Associates 1059 brief, frozen, pulverized tissue was lysed in chloroform/ Fisher Scientific). Peptides were separated on an easy-spray methanol (2:1) using a Tissue Lyser (Qiagen) at 30 Hz for column (500 mm 3 75 mm) with a flow rate of 250 nL/min 1 min, incubated for 20 min at 20°C with vigorous shaking and a gradient of 2–35% acetonitrile in 5% DMSO/0.1% (1,400 rpm; ThermoMixer; Eppendorf), and centrifuged at formic acid within 60 min. Detailed MS instrument settings 13,000 rpm for 30 min at 20°C. The liquid phase was mixed are listed in Supplementary Table 1. with 0.9% NaCl and centrifuged at 2,000 rpm for 5 min. LC-MS/MS data were analyzed using label-free precur- Chloroform/Triton X-100 (1:1) was added to the organic sor quantitation in Progenesis QI (version 3.0.6039.34628; phase and solvent evaporated under the fume hood. Trigly- Waters) and peptides identified with Mascot v2.7 (Matrix cerides were measured using the ab65336 kit (Abcam) and Science) against the UniProt/Swiss-Prot database (retrieved values normalized to the tissue pellet weight. 26 November 2015). Peptide false discovery rate was ad- justed to 1%, and additionally, all spectra identified with Glycogen ascore,20 were discarded. Glycogen was extracted from frozen, pulverized tissue samples using 30% KOH and a Tissue Lyser (30 Hz for 1 min; Metabolomics Qiagen), followed by incubation at 95°C for 30 min. Glycogen Mice were infused with two boluses of 20% U-13C-glucose was precipitated by addition of 95% ethanol and centrifu- (Cambridge Isotope Laboratories) in saline by intraperito- gation at 3,000g for 20 min. Pellets were dissolved in water neal injection at 2 mg U-13C-glucose/g mouse weight at 0 and and digested with amyloglucosidase (Sigma-Aldrich). Glucose 15 min. Hearts were harvested at 30 min post–first bolus was measured using a glucose (HK) kit (Sigma-Aldrich) and and immediately quenched by freeze-clamping at 280°C. normalized to protein content (Pierce BCA protein assay; Frozen tissue was lyophilized and pulverized. Powdered Thermo Fisher Scientific). material was transferred to a 1.5-mL Eppendorf tube and masses recorded. A total of 600 mL chloroform/methanol Western Blotting (2:1 volume for volume) was added to each sample and Proteins were extracted from frozen, pulverized tissue sam- vortexed briefly before pulse sonication (3 3 8min)in ples in a Tissue Lyser (30 Hz for 1 min; Qiagen) using a water-bath sonicator at 4°C for 1 h. Samples were spun a protein lysis buffer containing 50 mmol/L Tris, 150 mmol/L (13,200 rpm, 4°C, 10 min), supernatant transferred to a new NaCl, 1 mmol/L EDTA, 10 mmol/L NaF, 2 mmol/L Na3VO4, tube, and dried in a rotary vacuum concentrator. The 1 mmol/L dithiothreitol, 13 protease inhibitor cocktail remaining pellet was re-extracted with 600 mLmethanol/ (Sigma-Aldrich), and 1% Nonidet P-40. Protein concentra- water (2:1 volume for volume, containing 25 nmol nor- tion was measured using a Pierce BCA protein assay kit Leucine [internal standard 1]) followed by pulse sonication (Thermo Fisher Scientific). Proteins were separated on 4– for 8 min at 4°C. Samples were spun (as above) and the 12% Bis-Tris SDS-PAGE gels (NuPAGE Novex; Invitrogen), supernatant added to the first extract and dried. Extracts transferred to nitrocellulose membranes, and detected using were resuspended in 350 mL chloroform/methanol/water specific antibodies to PDK4 (8), medium-chain acyl-CoA (1:3:3) to partition polar and apolar metabolites. A total of dehydrogenase (MCAD; sc-49047; Santa Cruz Biotechnol- X mL of aqueous (polar) phase was dried with 1 nmol scyllo- ogy), UCP3 (ab3477; Abcam), HSC70 (ab19136; Abcam), and inositol (internal standard 2, X = volume of extract equiv- vinculin (ab129002; Abcam). For assessment of protein alent to 1 mg dry weight of heart tissue). Samples were oxidation, we used an Oxidized Protein Western Blot kit washed twice with methanol and derivatized with 20 mL (ab178020; Abcam). Proteins were quantified using ImageJ methoxyamine HCl (20 mg/mL in pyridine, overnight, at (National Institutes of Health). room temperature) and 20 mL BSTFA + TMCS (Thermo Fisher Scientific) for .1h. Protein Carbonylation Pulverized, frozen heart tissue was lysed in water, extracted, Gas Chromatography-MS and processed as recommended using a Protein Carbonyl Metabolite analysis was performed using gas chromatography- Content Assay kit (ab126287; Abcam). MS (GC-MS) using a 7890B-5977A system (Agilent Tech- nologies). Splitless injection (injection temperature 270°C) Proteomics ontoa30m+10m3 0.25 mm DB-5MS+DG column Samples were prepared for liquid chromatography–tandem (Agilent Technologies) was used, with helium as the carrier mass spectrometry (LC-MS/MS) analysis as described (23). gas, in electron impact ionization mode. The initial oven Briefly, proteins were precipitated with chloroform/methanol temperature was 70°C (2 min), followed by temperature following cell lysis in radioimmunoprecipitation assay gradients to 295°C at 12.5°C/min and then to 320°C at buffer and reduction/alkylation with dithiothreitol/iodoace- 25°C/min (held for 3 min). Metabolites were identified and tamide. After protein digestion with trypsin, peptides were quantified by comparison with the retention times and purified on reverse-phase material (SOLA SPE; Thermo mass spectra of authentic standards using MassHunter Fisher Scientific) and injected into a nano–LC-MS/MS work- WorkStation software (B.06.00 SP01; Agilent Technologies). flow consisting of a Dionex Ultimate 3000 UPLC and an Label incorporation and abundance was estimated using Orbitrap Fusion Lumos instrument (both from Thermo GAVIN software (24). Label incorporation was insignificant 1060 Cardiac Dysfunction in a Eulipidemic Model of Diabetes Diabetes Volume 67, June 2018 in diabetic mice due to their high level of (unlabeled) blood TTTCAGAGCATTGGCCATAGAA-39; and Rsp18 forward, glucose. 59-TGTGTTAGGGGACTGGTGGACA-39 and Rsp18 reverse, 9 9 DNA Isolation 5 -CATCACCCACTTACCCCCAAAA-3 . DNA was extracted from frozen, pulverized tissue samples Statistics using a DNeasy Blood and Tissue kit (Qiagen) following the Data are given as mean 6 SEM of the indicated number of ’ manufacturer sinstructions. mice (n). Significance was tested by t test or one-way or two- Quantitative PCR way ANOVA as indicated. Where data were found not to be RNA was extracted from frozen, pulverized tissue samples normally distributed, nonparametric tests were used (i.e., using Qiazol lysis reagent and a Tissue Lyser (30 Hz, 1 min), Kruskal-Wallis test). Differences between groups were con- followed by an RNeasy Mini kit (all from Qiagen). For sidered statistically significant if P was ,0.05. determination of expression of mitochondrially encoded genes, we included an on-column DNase digestion step RESULTS (Qiagen). cDNA was transcribed using the High Capacity Transgene induction in adult bV59M mice (15,16) led to cDNA Reverse Transcription kit (Applied Biosystems). marked hyperglycemia (blood glucose .20 mmol/L) within Quantitative PCRs were performed using TaqMan probes 24 h, and serum glucose remained elevated over the next (Supplementary Table 2) and the fast reaction kit on the 12 weeks (Supplementary Fig. 1A and C). Serum insulin StepOnePlus instrument (Life Technologies). Data were levels were significantly reduced after 2 weeks and remained quantified according to the delta threshold cycle method so over the next 20 weeks of diabetes (Fig. 1A). Cardiac (25) and normalized to levels of actin RNA. glycogen levels were unaltered (Supplementary Fig. 1B). We For determination of genomic and mitochondrial DNA, detected no changes in serum triglyceride, FA, or cholesterol we performed a semiquantitative SYBR PCR using the levels or lipid accumulation in the liver or heart after 2 weeks Power SYBR Green PCR Master Mix (Applied Biosys- of diabetes (Table 1 and Fig. 1B). Serum triglyceride levels tems) and the following primers: mtCox2 forward, 59- were slightly elevated after 4 and 20 weeks of diabetes, hence ATAACCGAGTCGTTCTGCCAAT-39 andmtCox2reverse,59- we focused on studying 2-week diabetic mice.

Figure 1—Diabetes without dyslipidemia alters cardiac function and metabolism. A: Serum insulin levels of uninjected (Ctrl) and tamoxifen-injected control(Tx)miceandofbV59M mice diabetic for 24 h or 2, 4, or 20 weeks (w) (as indicated) or diabetic for 2 weeks followed by glibenclamide therapy for 2 weeks (Glib). Data points indicate individual mice; bars are the mean 6 SEM of n =6–11 animals. **P , 0.01; ***P , 0.001 (one-way ANOVA). B: Serum triglyceride levels of same mice as in A. C–E:Strokevolume(C), cardiac output (D), and PDH flux (E) measured under control conditions before tamoxifen injection (Ctrl), following 2 weeks of diabetes (2w), and after 2 weeks of diabetes followed by 2 weeks of glibenclamide therapy (Glib). Where possible, the same mouse was studied sequentially in all three conditions. Stroke volume was measured as the difference between the end diastolic lumen and end systolic lumen. Cardiac output was obtained from cine MR images. PDH flux was measured in HP-MRS as the hyperpolarized [1-13C]pyruvate, bicarbonate/pyruvate ratio, and spectra summed over 30 s, beginning at the first appearance of the pyruvate peak. For E, one data point of Glib is outside of the axis limit. Data points indicate individual animals. Bars indicate mean 6 SEM. *P , 0.05; **P , 0.01 (Mann–Whitney test). diabetes.diabetesjournals.org Rohm and Associates 1061

Table 1—Phenotypic characteristics of bV59M mice Ctrl Diabetes Glib Body weight (g) 26.3 6 1.2 26.2 6 0.6 26.9 6 1.1 Bloodglucose,randomfed(mmol/L) 7.06 0.2 25.5 6 2.5*** 7.8 6 0.6 Bloodglucose,fasted(mmol/L) 3.66 0.3 21.1 6 1.8*** 3.8 6 0.5 Serum FFA (mmol/L) 0.48 6 0.05 0.41 6 0.06 0.41 6 0.07 Serum (V)LDL cholesterol (mg/mL) 0.17 6 0.01 0.16 6 0.01 0.15 6 0.01 Liver triglycerides (nmol/mg wet tissue) 1.74 6 0.19 1.64 6 0.44 1.65 6 0.34 Heart triglycerides (nmol/ng pellet) 0.81 6 0.09 0.72 6 0.07 NA Phenotypic characteristics of bV59M mice before tamoxifen injection (Ctrl), 2 weeks after tamoxifen injection (Diabetes), and after 2 weeks of diabetes followed by 2 weeks of glibenclamide (Glib) treatment (2 3 25 mg/21-day release pellets). Data are mean 6 SEM; n =4–14 mice. NA, not available. ***P , 0.001 compared with Ctrl (Mann–Whitney test).

Glibenclamide therapy (15) reversed the hyperglycemia decreased pyruvate flux through PDH, resulting in “pooling” within 2 days (Supplementary Fig. 1C). Two weeks later, of these metabolites. Alternatively, the TCA cycle may switch free-fed and fasted plasma glucose and serum insulin levels to being fueled by FA b-oxidation (for acetyl-CoA) and/or were normal. Serum and organ lipids remained unaltered glutaminolysis (for carbon skeletons). In support of this, (Fig. 1A and B and Table 1). hyperglycemia also reduced the relative abundance of some (but not all) longer chain FA, including C14:0, C16:1, and Hyperglycemia Impairs Cardiac Function and C18:1a (Fig. 2F and Supplementary Fig. 3). Taken together, Metabolism these data argue that chronic hyperglycemia/hypoinsuline- fi No signi cant change in cardiac size was seen with 2 weeks mia induces a metabolic switch from oxidative metabolism of – of diabetes (Supplementary Fig. 2A D), but there was a sig- glucose to lactate production and a concomitant increase in fi ni cant decrease in both stroke volume and cardiac output FA b-oxidation. (Fig. 1C and D). This was associated with a marked change in Ideally, to assess fluxes, it would be valuable to measure cardiac metabolism, as measured by both hyperpolarized the amount of each metabolite that derived from 13C-labeled MRS (HP-MRS) (Fig. 1E) and metabolomics (Fig. 2). glucose. This was possible in wild-type animals but not in Bicarbonate production from hyperpolarized diabetic mice because their high blood glucose levels meant 13 [1- C]pyruvate decreased in hearts of diabetic mice (Fig. that the percentage of labeled glucose was too small to be fl 1E), indicating reduced pyruvate ux though PDH (9). Nei- resolved. ther lactate nor alanine production from [1-13C]pyruvate were significantly altered, although there was a trend toward Hyperglycemia Drives Changes in Proteins Involved in an increase in lactate (Supplementary Fig. 2E and F). Gliben- Glucose and FA Metabolism clamide therapy (2 weeks) completely reversed the impair- We next explored if the switch in cardiac metabolism in- ment of cardiac function and pyruvate metabolism induced duced by diabetes was due to changes in the levels of by 2 weeks of hyperglycemia (Fig. 1C–E and Supplementary metabolic proteins by performing global protein expression Fig. 2). profiling of hearts from control mice and mice that had been Diabetes also affected metabolite abundance in the heart diabetic for 24 h or 2 weeks (Fig. 3). A total of 2,449 grouped when measured in mice injected with U-13C-glucose using proteins at 1% false discovery rate were identified (2,228 GC-MS. There was a significant increase in pyruvate abun- quantified). Of these, 299 proteins exhibited differential dance (Fig. 2A), indicating either enhanced pyruvate pro- expression between at least two conditions (P , 0.05 by duction or reduced consumption. In light of the HP-MRS ANOVA) (Supplementary Table 3). data (Fig. 1E), the latter is the more likely. Lactate and Consistent with the HP-MRS and metabolomics data, alanine were also increased (Fig. 2A), again indicative of diabetes increased the abundance of proteins involved in FA glycolytic flux being diverted from the tricarboxyclic acid metabolism and altered the abundance of many proteins (TCA) cycle into lactate production. As expected, glucose was involved in glucose oxidation (Fig. 3 and Supplementary increased, and there was a trend toward an increase in other Table 3). glycolytic intermediates (Fig. 2C). Fructose was strongly PDK4 was the most strongly upregulated protein in the elevated (Fig. 2D), as were the branched-chain amino acids glucose oxidation protein cluster (4.8-fold) (Fig. 3). PDK4 (BCAA) isoleucine and valine (Fig. 2E). Somewhat counter- regulates the activity of PDH by phosphorylating and inac- intuitively, and despite the unaltered abundance of many tivating the enzyme. Its upregulation may explain why PDH other metabolites (Supplementary Fig. 3), the TCA cycle flux is reduced in diabetes. Two key enzymes in the TCA metabolites citrate, cis-aconitate, succinate, and fumarate cycle, citrate synthase and oxoglutarate dehydrogenase were significantly increased (Fig. 2B). This may reflect (OGDH), were downregulated. The latter is of special sig- a relative reduction in TCA turnover as a consequence of nificance as it is the rate-limiting enzyme of the TCA cycle, 1062 Cardiac Dysfunction in a Eulipidemic Model of Diabetes Diabetes Volume 67, June 2018

Figure 2—The diabetic heart displays differences in metabolite proﬁle. Metabolite abundance estimated by GC-MS expressed as nanomoles per milligram of lyophilized heart tissue (for polar metabolites) or as relative abundance (for FA) in control (Ctrl) mice (black) and 2-week diabetic (2wdb) mice (white). Mean 6 SEM. Panels depict lower glycolysis (A), TCA cycle intermediates (B), upper glycolysis (C), fructose (D), BCAA (E), and FA (F). Nomenclature for FA: Cx:y, where x = number of carbons and y = number of double bonds, and “a” and “b” sufﬁxes indicate two FA differing only in double-bond position. *P , 0.05; **P , 0.01; ***P , 0.001 (t test). F6P, fructose 6-phosphate; G6P, glucose 6-phosphate; PEP, phosphoenol- pyruvate.

and its downregulation will reduce TCA cycle activity. The b-enolase (ENO3), and pyruvate kinase, was downregulated. defect in TCA cycle activity was not associated with altered Lactate dehydrogenase (LDHB) and the lactate transporter mitochondrial number (as measured by mitochondrial vs. (MCT1) were also reduced, despite the elevated lactate genomic DNA content and expression of mitochondrially production. encoded genes) (Supplementary Fig. 4A and B)norwith Notwithstanding the fact that the heart is not considered enhanced protein oxidation/carbonylation (Supplementary to be a ketogenic organ, the most strongly elevated protein Fig. 4C and D). Expression of several enzymes involved in the involved in lipid metabolism was 3-hydroxy-3-methylglutaryl- late steps of glycolysis, including phosphoglycerate kinase 1, CoA synthase 2 (mitochondrial, 21.4-fold), which catalyzes diabetes.diabetesjournals.org Rohm and Associates 1063

Interestingly, proteins involved in contractile function (such as myoglobin, a-actinin 2, and myosin 6) were downregulated in diabetic hearts (Supplementary Table 3), as were proteins involved in the creatine kinase shuttle (e.g., CKMT2 and CKB). In summary, the proteomics data support the metabolic data in suggesting glucose oxidation is impaired and FA metabolism enhanced. In line with the proteomics data, diabetes dramatically increased PDK4 mRNA (Fig. 4A) and protein expression (Fig. 4B). PDK4 is transcriptionally regulated by peroxisome proliferator–activated receptor a (PPARa), which is upregulated in the hearts of many rodent models of diabetes (26–28). However, Ppara mRNA expression in bV59M mouse heart was unchanged by hyperglycemia (Fig. 4C). Similarly, Pgc1a mRNA, which is elevated in other mouse models of diabetes (26,27) and regulates both PDK4 expression and PDH activity (29), was unaltered (Fig. 4D). Expression of the GLUT Glut4 decreased in bV59M diabetic mouse hearts (Fig. 4E). MCAD and UCP3 are major regulators of FA usage in theheart(9,30),andtheirexpressionincreasesinheartsof diabetic rodents with dyslipidemia (9). However, MCAD mRNA and protein were unaltered by diabetes in bV59M hearts (Fig. 5A and C and Supplementary Fig. 4E). This is consistent with the unchanged levels of cardiac lipids (Table 1) and Ppara/Pgc1a mRNA expression. Although PPARa is also regulated posttranscriptionally by ligand (lipid) binding, the lack of an increase in expression of its target gene (MCAD) suggests PPARa is not a major transcriptional regulator in this setting. Both UCP3 mRNA and protein increased significantly (Fig. 5B and C and Supple- mentary Fig. 4). The increase in Ucp3 and Pdk4 mRNA was observed as early as 24 h after gene induction, and elevated levels persisted over the next 20 weeks (Fig. 6A and B). At the protein level, increased amounts of PDK4 and UCP3 were only detectable after 2 and 4 weeks of diabetes, respectively (Fig. 6C). However, a reduction in cardiac output and stroke volume was observed within 2 weeks of diabetes induction (Figs. 1C and D and 6D and E).Prolongeddiabetes(8weeks) led to further deterioration of cardiac function, with both the ejection fraction and heart rate also being reduced (Fig. Figure 3—Hyperglycemia-induced changes in protein expression. 6F and G). Changes in abundance of selected proteins involved in lipid and carbohydrate metabolism. Expression in hearts of tamoxifen-injected control(Ctrl)miceandbV59M mice diabetic (db) for 24 h or 2 weeks (w). Each lane indicates a different mouse. Levels are compared with the DISCUSSION mean of that found for control mice and depicted as a log(2)-fold Using a combination of HP-MRS, metabolomics, MRI, and change spanning from 21 in dark blue to +1 in dark red (as indicated). See Supplementary Table 3 for full data. proteomics, we show that diabetes (hyperglycemia/hypoinsulinemia) drives changes in cardiac metabolism independent of (or in addition to) plasma FA/lipids. In particular, it the first step in ketogenesis. The mitochondrial carnitine/ reduces flux through PDH and thereby impairs entry of acylcarnitine carrier protein SLC25A20, which mediates FA pyruvate into the TCA cycle. Although mitochondrial num- import into mitochondria for oxidation, was also increased bers seem unaltered, the relative amount of glucose oxidized (2.1-fold). In addition, many proteins involved in FA oxi- by the mitochondria is reduced, leading to the observed dation (e.g., ACOX1 and SCP2) or lipolysis (e.g., PLIN5 and increase in lactate production. Proteins involved in FA MGL) were elevated (Fig. 3). metabolism increase, and long-chain FA levels decrease, 1064 Cardiac Dysfunction in a Eulipidemic Model of Diabetes Diabetes Volume 67, June 2018

Figure 4—PDK4 expression is induced in diabetes. A: Pdk4 mRNA expression in hearts of noninjected control mice (Ctrl; n =6),bV59M mice after 4 weeks of diabetes (4w db; n =9),andbV59M mice that were diabetic for 2 weeks and then treated with glibenclamide for 2 weeks (Glib; n =7). Data points indicate individual animals; bars indicate mean 6 SEM. B: PDK4 and vinculin protein levels in hearts of control (Ctrl), 2-week diabetic bV59M mice (2w db), and 2-week diabetic plus 2-week glibenclamide-treated bV59M mice (Glib). Each lane is a separate mouse. Ppara (C), Pgc1a (D), and Glut4 (E)mRNAexpressioninheartsofthesamemiceasinA.***P , 0.001 (one-way ANOVA).

indicating FA metabolism is enhanced. These alterations in changes we observe. First, serum insulin is reduced but not cardiac metabolism lead to impaired cardiac function, as in- absent in bV59M mice (15) (Fig. 1A). Second, preventing dicated by the reduction in cardiac output and stroke volume. hyperglycemia using dapagliflozin preserved cardiac function We found that changes in cardiac metabolism manifest very in a mouse model of lipodystrophy (11). Third, deletion of rapidly: PDK4 expression is already altered within 24 h of the insulin receptor specifically in cardiac myocytes (pre- diabetes induction. Thus, glucose is a very fast driver of cardiac venting insulin action) produced very different changes in dysfunction. Indeed, although some studies found increased gene expression and metabolism to those seen in bV59M PDK4 levels already after 24 h of high-fat diet feeding (31), mice (32). For example, PDK4 and MCAD expression were others show that high-fat feeding only impairs cardiac metabo- downregulated, GLUT4 expression was upregulated, and FA lism and function after a longer exposure (30). We also dem- oxidation was reduced; these changes are all the opposite of onstrate that the changes in cardiac metabolism and function those seen in our mice. Nevertheless, although we favor the induced by diabetes are reversed when euglycemia is restored. idea that chronic hyperglycemia is the primary cause of many Importantly, because diabetic bV59M mice are not obese of the changes we observe, we cannot exclude a role for and show no changes in serum or tissue lipids at the time hypoinsulinemia. Diabetes is a multifactorial disease, and both point studied, the effects we observe are the result of hyperglycemia and hypoinsulinemia are important. We can, hyperglycemia/hypoinsulinemia and not obesity or dys- however, exclude a role for serum dyslipidemia in our model. lipidemia. Our findings are supported by a recent study of As previously reported for other diabetes models (33), 2 2 the Bscl2 / mouse model of lipodystrophy (11). These mice diabetic bV59M mice had lower expression of Glut4 in the exhibited cardiac dysfunction without any change in intra- heart. Nevertheless, cytosolic glucose levels were slightly myocardial lipids, which was suggested to result from elevated elevated in diabetes, presumably due to the marked increase plasma glucose, as it could be reversed by the glucose-lowering in blood glucose. This suggests the decrease in GLUT4 is sodium–glucose cotransporter 2 inhibitor dapagliflozin. a protective mechanism, producing a relative reduction in The hyperglycemia of bV59M mice results from hypo- glucose uptake in the face of chronic hyperglycemia. insulinemia. Several lines of evidence suggest that it is primarily Many proteins involved in glucose metabolism, such hyperglycemia and not hypoinsulinemia that drives the cardiac as pyruvate kinase, citrate synthase, and OGDH, were diabetes.diabetesjournals.org Rohm and Associates 1065

ketogenesis (e.g., 3-hydroxy-3-methylglutaryl-CoA synthase 2), b-oxidation (ACOX1 and SCP2), mitochondrial FA import (SLC25A20), and lipolysis (PLIN5 and MGL). These changes suggest enhanced b-oxidation in hyperglycemia. Metabolo- mics analysis supports this idea, revealing a marked decrease in myristic (C14:0) and palmitoleic (C16:1) acids. Thus, our data suggest that hyperglycemia/hypoinsulinemia leads to a paradoxical increase in FA usage and a concomitant decrease in glucose usage in the diabetic heart, perhaps con- tributing to TCA flux maintenance. Metabolomics analysis demonstrated a marked increase in the BCAA isoleucine and valine in diabetic hearts. It is possible this reflects plasma levels of these substances, which are enhanced in diabetes (37). However, the heart itself catab- olizes BCAA and expresses high levels of enzymes involved in BCAA oxidation (38), such as BCAA aminotransferase and branched-chain keto acid dehydrogenase (BCKDHA and BCKDHB). These enzymes are also highly expressed in the mouse heart, as detected by proteomics, but they are not significantly altered by diabetes (Supplementary Table 3). It is also noteworthy that inhibition of cardiac BCAA catabolism in mice leads to decreased PDH activity and glucose oxidation (39). Figure 5—UCP3 levels are increased by hyperglycemia. Mcad (A)and PDH is the key enzyme regulating substrate usage in the Ucp3 (B) mRNA levels in control mice (Ctrl; n =6)andbV59M mice after heart. Reduced PDH activity decreases mitochondrial car- 4weeksofdiabetes(4wdb;n = 9) or 2 weeks of diabetes followed by bohydrate metabolism and enhances lactate production. 2 weeks of glibenclamide therapy (Glib; n = 7). Data points indicate individual animals; bars are mean 6 SEM. C: UCP3, MCAD, and This favors FA usage, impairing cardiac energy production HSC70 (loading control) protein levels in control mice (Ctrl) and and thereby cardiac function (40). Cardiac PDH activity is bV59M mice after 2 weeks of diabetes (2w db) or 2 weeks of diabetes decreased in many animal models of diabetes (9,41,42), most followed by 2 weeks of glibenclamide therapy (Glib). Each lane is of which are both hyperglycemic and dyslipidemic, and a different mouse. **P , 0.01; ***P , 0.001 (one-way ANOVA). cardiac Pdk4 mRNA levels are dramatically increased in streptozotocin-injected diabetic rats (43). As we show in this study, hyperglycemia/hypoinsulinemia is sufficient to decreased in diabetes. OGDH is a subunit of the enzyme reduce PDH activity, likely via transcriptional upregulation complex that converts 2-oxoglutarate to succinyl-CoA and of PDK4. This does not exclude a role for FA in regulating CO2. This is the rate-limiting reaction in the TCA cycle (34), PDK4 expression. Indeed, PDK4 expression is induced by and a reduction in OGDH, together with the reduced PDH long-chain FA (e.g., palmitate) (44), and both increased Pdk4 flux, may lead to impaired ATP generation by diabetic hearts, mRNA and reduced PDH flux were found in hearts of rats causing functional defects. Mitochondrial dysfunction and fed a high-fat diet (9). Thus, an additional contribution from reduced glucose oxidation have been reported in mitochon- elevated lipids may occur when hyperglycemia is associated dria isolated from hearts of donors with T2D (4). with obesity and dyslipidemia. Interestingly, inhibition of PDK It was somewhat surprising that we saw no increase in activity by the pyruvate mimetic dichloroacetate normal- protein carbonylation in diabetic hearts, as it is well estab- ized cardiac metabolism and diastolic function in high- lished that cardiac oxidative stress occurs in diabetes (35). fat–fed rats, arguing that reduced PDH flux contributes to The extent of carbonylation is considered a marker for diabetes-associated cardiac dysfunction (9). oxidative stress, with the advantage that it is relatively stable The major transcription factor regulating PDK4 in the and induced by almost all types of reactive oxygen species heart is thought to be PPARa (9). It also elevates MCAD and (36). It is possible that our failure to detect protein carbon- UCP3 expression (9). Although PPARa expression is induced ylation indicates that 2–4 weeks of diabetes is insufficient in diabetes (26–28), PPARa is primarily regulated posttran- to induce measurable oxidative stress, it is insufficient for scriptionally (e.g., by lipids [45]). There was little change in oxidative stress to produce protein carbonylation, or oxida- intracellular lipids and Ppara mRNA expression in diabetic tive stress is less when hyperglycemia/hypoinsulinemia is bV59M hearts. Furthermore, the lack of a change in MCAD not accompanied by dyslipidemia. It may, however, explain expression argues PPARa activity may also be unchanged why we observe a near-complete reversibility of the cardiac and that other factors may underlie the marked increase in phenotype upon restoration of euglycemia. PDK4 expression. In this context, it is important to recall Many proteins involved in lipid metabolism were upre- that PPAR response elements are absent in the promoter gulated in diabetes. These included proteins involved in region of PDK4 (46). 1066 Cardiac Dysfunction in a Eulipidemic Model of Diabetes Diabetes Volume 67, June 2018

Figure 6—Time-dependent deterioration of cardiac function and metabolism. Pdk4 (A)andUcp3 (B) mRNA levels in uninjected (Ctrl) and tamoxifen-injected control (Tx) mice and in bV59M mice diabetic for 24 h or 2, 4, or 20 weeks (w) (as indicated) or diabetic for 2 weeks followed by glibenclamide therapy for 2 weeks (Glib). Levels are expressed relative to Ctrl mice. Data points indicate individual mice; bars are mean 6 SEM of n =6–9animals.C: Protein levels of PDK4, UCP3, and HSC70 (loading control) in the same mice as in A and B. Each lane is a different mouse. Cardiac output (D), stroke volume (E), ejection fraction (F), and heart rate (G) measured by cine MRI in the same bV59M mice (n =7–15) ﬁrst under control conditions (Ctrl) and then following 2 (2w) or 8 weeks (8w) of diabetes. Data points indicate individual animals; bars are mean 6 SEM. *P , 0.05; **P , 0.01; ***P , 0.001 (one-way ANOVA). BPM, beats per minute.

Our results demonstrate that hyperglycemia/hypoinsuli- performed in the Target Discovery Institute Mass Spectrometry Laboratory led by nemia alone is sufﬁcient to impair cardiac function, that this Benedikt M. Kessler. can occur very rapidly following elevation of blood glucose, Funding. This work was supported by the Wellcome Trust (grants 084655 and and that (at least after diabetes of short duration) it may be 089795) and the European Research Council (ERC advanced grant 322620). M.R. was supported by a Novo Nordisk postdoctoral fellowship run in partnership with the reversed by restoration of euglycemia. This suggests that University of Oxford. R.F. was supported by the Kennedy Trust Fund. Metabolomics even nonobese patients with diabetes, or those without was supported by The Francis Crick Institute, which receives its core funding from systemic hyperlipidemia, may already have some cardiac Cancer Research UK (FC001999), the Medical Research Council (FC001999), and the impairment at presentation. It also emphasizes the need Wellcome Trust (FC001999). F.M.A. held a Royal Society/Wolfson Merit Award and a for tight control of blood glucose levels, which may not only European Research Council Advanced Investigatorship. The HP-MRS studies were prevent but also ameliorate the deleterious effects of di- supported by the British Heart Foundation (Fellowships FS/10/002/28078 and abetes on the heart. FS/14/17/30634), the Oxford British Heart Foundation Centre of Research Excellence (grant RE/13/1/30181), and the Danish Council for Strategic Research (LIFE-DNP Programme). Acknowledgments. The authors thank Lisa Heather, Duncan Sparrow, Mike Duality of Interest. No potential conﬂicts of interest relevant to this article Dodd (Department of Physiology, Anatomy and Genetics, University of Oxford), and were reported. Markus Ralser (The Francis Crick Institute, London) for helpful discussion and Raul Author Contributions. M.R. performed the animal and molecular biology Terron Exposito, Olof Rorsman, and Idoia Portillo (Department of Physiology, Anatomy work and analyzed the data. M.R. and F.M.A. designed the study and wrote the and Genetics, University of Oxford) for technical support. Proteomics analysis was manuscript. D.S., V.B., and M.K.C. performed the cine MRI and HP-MRS diabetes.diabetesjournals.org Rohm and Associates 1067

measurements. D.S. and D.J.T. analyzed the cine MRI and HP-MRS data. S.B. and 23. Lochmatter C, Fischer R, Charles PD, Yu Z, Powrie F, Kessler BM. Integrative R.F. performed proteomics and analyzed the data. N.L. and J.I.M. performed phosphoproteomics links IL-23R signaling with metabolic adaptation in lym- metabolomics and analyzed the data. All authors read the manuscript and phocytes. Sci Rep 2016;6:24491 discussed the interpretation of the results. D.J.T. and F.M.A. are the guarantors 24. Behrends V, Tredwell GD, Bundy JG. A software complement to AMDIS for of this work and, as such, had full access to all of the data in the study and take processing GC-MS metabolomic data. Anal Biochem 2011;415:206–208 responsibility for the integrity of the data and the accuracy of the data analysis. 25. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001;25:402–408 References 26. Duncan JG, Fong JL, Medeiros DM, Finck BN, Kelly DP. Insulin-resistant heart 1. Kannel WB, McGee DL. Diabetes and cardiovascular risk factors: the Fra- exhibits a mitochondrial biogenic response driven by the peroxisome proliferator- – mingham study. Circulation 1979;59:8 13 activated receptor-alpha/PGC-1alpha gene regulatory pathway. Circulation 2007; 2. Jia G, Whaley-Connell A, Sowers JR. Diabetic cardiomyopathy: a hyperglycaemia- 115:909–917 – and insulin-resistance-induced heart disease. Diabetologia 2018;61:21 28 27. Mitra R, Nogee DP, Zechner JF, et al. The transcriptional coactivators, PGC-1a 3. Scheuermann-Freestone M, Madsen PL, Manners D, et al. Abnormal cardiac and b, cooperate to maintain cardiac mitochondrial function during the early stages and skeletal muscle energy metabolism in patients with type 2 diabetes. Circulation of insulin resistance. J Mol Cell Cardiol 2012;52:701–710 – 2003;107:3040 3046 28. Finck BN, Lehman JJ, Leone TC, et al. The cardiac phenotype induced by 4. Montaigne D, Marechal X, Coisne A, et al. Myocardial contractile dysfunction PPARalpha overexpression mimics that caused by diabetes mellitus. J Clin Invest is associated with impaired mitochondrial function and dynamics in type 2 diabetic 2002;109:121–130 – but not in obese patients. Circulation 2014;130:554 564 29. Kiilerich K, Adser H, Jakobsen AH, et al. PGC-1alpha increases PDH content but 5. Anderson EJ, Kypson AP, Rodriguez E, Anderson CA, Lehr EJ, Neufer PD. does not change acute PDH regulation in mouse skeletal muscle. Am J Physiol Regul fi Substrate-speci c derangements in mitochondrial metabolism and redox Integr Comp Physiol 2010;299:R1350–R1359 balance in the atrium of the type 2 diabetic human heart. J Am Coll Cardiol 2009; 30. Cole MA, Murray AJ, Cochlin LE, et al. A high fat diet increases mitochondrial fatty 54:1891–1898 acid oxidation and uncoupling to decrease efficiency in rat heart. Basic Res Cardiol 6. Isfort M, Stevens SC, Schaffer S, Jong CJ, Wold LE. Metabolic dysfunction in 2011;106:447–457 diabetic cardiomyopathy. Heart Fail Rev 2014;19:35–48 31. Crewe C, Kinter M, Szweda LI. Rapid inhibition of pyruvate dehydrogenase: 7. Abel ED, O’Shea KM, Ramasamy R. Insulin resistance: metabolic mechanisms an initiating event in high dietary fat-induced loss of metabolic flexibility in the heart. and consequences in the heart. Arterioscler Thromb Vasc Biol 2012;32:2068–2076 PLoS One 2013;8:e77280 8. Zlobine I, Gopal K, Ussher JR. Lipotoxicity in obesity and diabetes-related 32. Belke DD, Betuing S, Tuttle MJ, et al. Insulin signaling coordinately regulates cardiac dysfunction. Biochim Biophys Acta 2016;1861:1555–1568 cardiac size, metabolism, and contractile protein isoform expression. J Clin Invest 9. Le Page LM, Rider OJ, Lewis AJ, et al. Increasing pyruvate dehydrogenase flux 2002;109:629–639 13 as a treatment for diabetic cardiomyopathy: a combined C Hyperpolarized Magnetic 33. Shao D, Tian R. Glucose transporters in cardiac metabolism and hypertrophy. Resonance and Echocardiography Study. Diabetes 2015;64:2735–2743 Compr Physiol 2015;6:331–351 10. Drosatos K, Schulze PC. Cardiac lipotoxicity: molecular pathways and 34. Shi Q, Xu H, Yu H, et al. Inactivation and reactivation of the mitochondrial therapeutic implications. Curr Heart Fail Rep 2013;10:109–121 a-ketoglutarate dehydrogenase complex. J Biol Chem 2011;286:17640–17648 11. Joubert M, Jagu B, Montaigne D, et al. The sodium-glucose cotransporter 35. Faria A, Persaud SJ. Cardiac oxidative stress in diabetes: mechanisms and 2 inhibitor dapagliflozin prevents cardiomyopathy in a diabetic lipodystrophic mouse therapeutic potential. Pharmacol Ther 2017;172:50–62 model. Diabetes 2017;66:1030–1040 36. Hecker M, Wagner AH. Role of protein carbonylation in diabetes. J Inherit Metab 12. Teleb M, Popp Switzer M, Elhanafi S, Elfar A, San Juan ZT. Glycemic control Dis 2018;41:29–38 and excess cardiovascular mortality in type 1 diabetes. Curr Cardiol Rep 2016;18:29 37. Xu F, Tavintharan S, Sum CF, Woon K, Lim SC, Ong CN. Metabolic signature 13. Aguilar-Bryan L, Bryan J. Neonatal diabetes mellitus. Endocr Rev 2008;29:265–291 shift in type 2 diabetes mellitus revealed by mass spectrometry-based meta- 14. Iribarren C, Karter AJ, Go AS, et al. Glycemic control and heart failure among bolomics. J Clin Endocrinol Metab 2013;98:E1060–E1065 adult patients with diabetes. Circulation 2001;103:2668–2673 38. Harper AE, Miller RH, Block KP. Branched-chain amino acid metabolism. Annu 15. Brereton MF, Iberl M, Shimomura K, et al. Reversible changes in pancreatic islet Rev Nutr 1984;4:409–454 structure and function produced by elevated blood glucose. Nat Commun 2014;5: 39. Li T, Zhang Z, Kolwicz SC Jr, et al. Defective branched-chain amino acid 4639 catabolism disrupts glucose metabolism and sensitizes the heart to ischemia- 16. Brereton MF, Rohm M, Shimomura K, et al. Hyperglycaemia induces metabolic reperfusion injury. Cell Metab 2017;25:374–385 dysfunction and glycogen accumulation in pancreatic b-cells. Nat Commun 2016;7: 40. Chatham JC, Forder JR. Relationship between cardiac function and substrate 13496 oxidation in hearts of diabetic rats. Am J Physiol 1997;273:H52–H58 17. Schneider JE, Cassidy PJ, Lygate C, et al. Fast, high-resolution in vivo cine 41. Hall JL, Stanley WC, Lopaschuk GD, et al. Impaired pyruvate oxidation but normal magnetic resonance imaging in normal and failing mouse hearts on a vertical 11.7 T glucose uptake in diabetic pig heart during dobutamine-induced work. Am J Physiol system. J Magn Reson Imaging 2003;18:691–701 1996;271:H2320–H2329 18. Haase A, Frahm J, Matthaei D, Hänicke W, Merboldt KD. FLASH imaging: rapid 42. Seymour AM, Chatham JC. The effects of hypertrophy and diabetes on cardiac NMR imaging using low flip-angle pulses. 1986. J Magn Reson 2011;213:533–541 pyruvate dehydrogenase activity. J Mol Cell Cardiol 1997;29:2771–2778 19. Ardenkjaer-Larsen JH, Fridlund B, Gram A, et al. Increase in signal-to-noise ratio 43. Wu P, Sato J, Zhao Y, Jaskiewicz J, Popov KM, Harris RA. Starvation and diabetes of . 10,000 times in liquid-state NMR. Proc Natl Acad Sci U S A 2003;100:10158– increase the amount of pyruvate dehydrogenase kinase isoenzyme 4 in rat heart. 10163 Biochem J 1998;329:197–201 20. Dodd MS, Ball V, Bray R, et al. In vivo mouse cardiac hyperpolarized magnetic 44. Huang B, Wu P, Bowker-Kinley MM, Harris RA. Regulation of pyruvate de- resonance spectroscopy. J Cardiovasc Magn Reson 2013;15:19 hydrogenase kinase expression by peroxisome proliferator-activated receptor-alpha 21. Vanhamme L, van den Boogaart A, Van Huffel S. Improved method for accurate ligands, glucocorticoids, and insulin. Diabetes 2002;51:276–283 and efficient quantification of MRS data with use of prior knowledge. J Magn Reson 45. Schilling JD. The mitochondria in diabetic heart failure: from pathogenesis to 1997;129:35–43 therapeutic promise. Antioxid Redox Signal 2015;22:1515–1526 22. Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation and 46. Yang J, Sambandam N, Han X, et al. CD36 deficiency rescues lipotoxic car- purification of total lipides from animal tissues. J Biol Chem 1957;226:497–509 diomyopathy. Circ Res 2007;100:1208–1217