Computational Assessment of Myocardial Metabolism in Ischaemic and Dilated Cardiomyopathy

J. Wang (Berlin)¹, N. Berndt (Nuthetal)², M. Kirchner (Berlin)³, V. Braun (Berlin)⁴, C. Knosalla (Berlin)¹, I. Baczko (Szeged)⁵, V. Falk (Berlin)⁶, P. Mertins (Berlin)³, F. Schönrath (Berlin)¹, H. Milting (Bad Oeynhausen)⁷, K. Beckschulte (Berlin)¹, N. Beyhoff (Berlin)⁴, L. Finnigan (Berlin)^1
1Deutsches Herzzentrum der Charite (DHZC) Klinik für Herz-, Thorax- und Gefäßchirurgie Berlin, Deutschland; ²Deutsches Institut für Ernährungsforschung Potsdam-Rehbrücke (DIfE) Molekulare Toxikologie Nuthetal, Deutschland; ³Max-Delbrück-Centrum für Molekulare Medizin Berlin, Deutschland; ⁴Charité - Universitätsmedizin Berlin Institut für Pharmakologie Berlin, Deutschland; ⁵University of Szeged Department of Pharmacology and Pharmacotherapy Szeged, Ungarn; ⁶Charité - Universitätsmedizin Berlin Klinik für kardiovaskuläre Chirurgie Berlin, Deutschland; ⁷Herz- und Diabeteszentrum NRW E.& H. Klessmann-Institut f. kardiovask. Forschung Bad Oeynhausen, Deutschland

Background:  While cardiac energetic impairment is a hallmark feature of heart failure, the impact of the underlying aetiology on cardiac metabolism remains poorly understood. This study aimed to define myocardial bioenergetic phenotypes in a multicentre cohort of patients with advanced ischaemic cardiomyopathy (ICM) and dilated cardiomyopathy (DCM) together with healthy controls using computational modelling.

Methods: A total of 118 myocardial tissue samples (DCM, n= 41; ICM, n= 34; controls, n= 43) underwent quantitative proteomic profiling using high-resolution mass spectroscopy. Patient-specific bioenergetic phenotypes were reconstructed by integrating the proteomics-derived enzyme abundance profile into a comprehensive kinetic model encompassing 296 biochemical reactions of myocardial energy metabolism. This computational framework enables simulations of changes in myocardial workload, substrate availability, and enzyme regulation. A general linear model was used to assess differences between ICM, DCM and control groups.

Results: Maximal ATP production capacity was lower in ICM (mean diff = 190.7; 95% CI = 70.0, 311.5; P < 0.001) and DCM hearts (mean diff = 169.4; 95% CI = 54.6, 284.3; P = 0.002) compared to controls. Similarly, maximal O₂ consumption was also lower in ICM (mean diff = 42.4; 95% CI = 13.7, 71.0; P = 0.002) and DCM (mean diff = 38.5; 95% CI = 11.3, 65.8; P = 0.003) when compared to the control group. Respiratory efficiency, the ATP yield per mole O₂, was preserved.

Maximal glucose uptake capacity was higher in DCM hearts compared to ICM (mean diff = 6.41; 95% CI = 0.32, 12.5; P = 0.04) and compared to controls (mean diff = 6.22; 95% CI = 0.44, 12.0; P = 0.03). The myocardial uptake capacities for fatty acids, ketones, lactate, and branched-chain amino acids as energy-delivering substrates did not differ between disease groups. Compared to controls, utilisation of glucose relative to fatty acids was higher in the ICM and DCM groups, depending on myocardial workload and substrate availability. This shift in myocardial substrate utilisation was more pronounced in DCM than in ICM (mean diff = 1.24; 95% CI = 0.06, 2.42; P = 0.04).

Conclusion: Computational modelling of myocardial metabolism identified distinct bioenergetic signatures in ICM and DCM despite comparable heart failure severity. This highlights the potential of systems-based metabolic phenotyping to potentially guide personalised diagnosis and targeted metabolic therapies.